Interhemispheric connectivity during spontaneous switches in visual perceptual coherence as revealed by fMRI at multiple temporal resolutions

Interhemispheric connectivity during spontaneous switches in visual perceptual coherence as revealed by fMRI at multiple temporal resolutions | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Interhemispheric connectivity during spontaneous switches in visual perceptual coherence as revealed by fMRI at multiple temporal resolutions Alexandre Sayal, Bruno Direito, Teresa Sousa, João Duarte, Sónia Afonso, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4335511/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Functional magnetic resonance imaging (fMRI) allows to observe neural activity in real-time but tracking the neural correlates of perceptual decision as a function of interhemispheric connectivity has remained difficult. Recent advances in image acquisition, namely with the surfacing of multiband sequences, have led us to investigate this mechanism using higher temporal resolution approaches. We were able to better capture the hemodynamic responses to rapid changes in neural activity concomitantly with a task requiring either perceptual interhemispheric segregation or integration, shortening the gap to other neuroimaging techniques, which is particularly significant when considering the study of dynamic connectivity patterns. Here, we tested the hypothesis whether interhemispheric connectivity in the visual cortex relates to interhemispheric integration, when presented with bistable moving stimuli at four distinct temporal resolutions. Based on this connectivity metric, we could discern perceptual state transitions related to connectivity. First, we found that activation response metrics to visual motion in our target region of interest, the human visual motion complex hMT+, are stable across temporal resolutions. Then, we investigated interhemispheric connectivity between homologous hMT + in response to bistable moving stimuli, for all resolutions, which was critical for replication of perception related interhemispheric synchrony. The established relation between perceptual coherence and increased synchrony across the hemispheres suggests the feasibility of a real-time fMRI neurofeedback based on interhemispheric connectivity. Accordingly, we could infer perceptual states based on this connectivity metric while designing a rule that could even be used to generate feedback. We further showed that higher resolution sequences are beneficial when implementing feedback interfaces based on interhemispheric functional connectivity, both regarding the delay and the accuracy of the feedback itself. Regarding the use of real time fMRI and neurofeedback strategies, higher resolution sequences are likely needed, when relying on connectivity metrics. Biological sciences/Neuroscience/Cognitive neuroscience/Perception Biological sciences/Neuroscience/Sensory processing Biological sciences/Neuroscience/Neural circuit Visual perception bistable motion interhemispheric connectivity neurofeedback multiband fMRI Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Introduction Functional magnetic resonance imaging (fMRI) has revolutionized our understanding of brain function, and it is now possible to observe neural activity in real-time. One critical aspect in the context of fMRI is its temporal resolution. Unlike other neuroimaging techniques, such as electroencephalography (EEG) or magnetoencephalography (MEG), fMRI has limited temporal resolution, meaning it may not have the sensitivity to capture the hemodynamic responses to rapid changes in neural activity. This aspect becomes particularly significant when considering the study of dynamic connectivity patterns in the human visual system. The perceptual integration of components of motion to generate a global pattern may require communication between different parts of the visual system (Duarte et al., 2017 ). In the case of bistable perception, the fast changes in interhemispheric connectivity inform us about the binding mechanisms of visual motion features (Sousa et al., 2019 ). While aiming for a novel intervention that target online perceptual learning mechanisms (Scharnowski et al., 2012 ), the detection in real-time of functional connectivity changes is key. Hence, in this study, we explore the interhemispheric connectivity in the visual cortex in response to bistable (coherent or incoherent) moving stimuli at different temporal resolutions, to validate and infer perceptual states based on this metric. Online tracking of perception might also help prompt the exploration of a real-time fMRI connectivity based neurofeedback protocol that aims at potentiating cognitive functions in an immersive therapeutic setting. Ultimately, this output of neuroscientific relevance to track interhemispheric integration and perceptual decision would also allows us to generalize the impact of temporal resolution in real time approaches such as neurofeedback paradigms. Our first goal is to test the hypothesis that the functional dynamics of the human motion complex (hMT+) across hemispheres reflects bistable global motion perception. This region, homologous to the monkey MT region, is a well-known region of the visual cortex that responds to motion in the visual field. Here, we follow previous evidence acquired at 7T that suggested, for the first time, that perceptual long-range integration of bistable motion is linked to the interhemispheric functional connectivity in hMT+ (Sousa et al., 2019 ). However, that finding remains the single one in the literature suggesting that interhemispheric perceptual integration is mirrored by enhanced functional connectivity across the hemispheres Specifically, the working hypothesis is that interhemispheric functional connectivity hMT+, at the scale of fMRI temporal resolution, reflects perceptual decision (eg. Coherence vs Incoherence). Using a different bistable stimulus (moving plaids, Fig. 1 ) and a 3T MRI scanner, we aimed to validate these findings at different fMRI temporal resolutions, which depend on the acquisition Repetition Time (TR). Our motivation lies in the ability to detect fast perceptual transitions, in the order of magnitude of the TR, to then explore its neuroscientific potential as well as the possibility of using this information in a neurofeedback setup. We use a moving plaid, which is a bistable visual stimulus that can be seen as two gratings sliding over each other and moving in different directions (incoherent motion) or as a single surface moving coherently – the two gratings are integrated perceptually into a single surface moving coherently in a direction intermediate to the directions of the individual gratings (coherent motion) (Sousa et al., 2018 , 2024 ; Sayal et al., 2020 ) (Fig. 1 ) Multiband echo-planar imaging (MB-EPI) sequences have revolutionized image sampling rates, offering a remarkable increase in temporal resolution. These use advanced pulse sequences and specialized MRI hardware to simultaneously excite and acquire data from multiple slices. By doing so, they can cover a larger volume of the brain or body in a shorter time compared to the sequential acquisition of slices. In clinical applications, these sequences show more detailed BOLD signals with superior temporal resolution and contrast-to-noise ratio, pivotal features for detecting nuanced responses of neuronal populations (Zhang et al., 2024 ). However, high multiband factors should be used with caution, as discussed in a recent editorial (Wall, 2023 ). Given the increasing application of MB sequences for fMRI protocols, a number of studies started to thoroughly evaluate how the BOLD signal was affected by this new acquisition technique (Chen et al., 2015 ; Todd et al., 2016 , 2017 ; Demetriou et al., 2018 ; Risk, Kociuba and Rowe, 2018 ; Jahanian et al., 2019 ; Bhandari et al., 2020 ; Renz et al., 2023 ). These studies compare quality and functional metrics for data acquired with different temporal resolutions (hence multiband factors). For instance, (Chen et al., 2015 ) results revealed notably higher BOLD information content when employing faster TRs between 300 ms and 600 ms, in contrast to a 2-second TR, indicating that faster TRs can capture more information per unit of time in task-based fMRI studies. Another study (Todd et al. 2016 ) evaluated BOLD sensitivity and false-positive activations. Based on the results, the authors make recommendations regarding the reconstruction method and the MB factor that allows for a combined low probability of false-positive activations (due to slice leakage) and high quality (fair spatial and temporal resolution) data. Their conservative recommendation limits the MB factor to 2 and in-plane acceleration to 2. In fact, the benefits of these sequences are not always clear. (Demetriou et al. 2018 ) found that multi-band protocols showed strong benefits for resting-state data analyses and Multi-Voxel Pattern Analyses (MVPA) but not so evident and consistent improvements for task-based paradigms and standard General Linear Model (GLM) approaches, linking these results to variations in temporal signal-to-noise-ratio (SNR) between sequences. (Srirangarajan et al., 2021 ) provided a cautionary note regarding the study of mesolimbic regions with multiband sequences, as the authors found reductions in temporal SNR that could account for impaired detection of task-related responses in these deeper regions in the brain. Neurofeedback (NF) experiments based on real-time fMRI have opened the possibility for individuals to self-regulate specific not only brain activation but now also connectivity patterns. This is achieved by interpreting a feedback signal and exploring the appropriate strategies for its modulation. Currently, the impact of several protocol parameters with special attention given to exploring the influence of different self-regulation strategies, feedback interfaces, and data analysis pipelines (Sitaram et al., 2017 ; Hampson, Ruiz and Ushiba, 2020 ; Ros et al., 2020 ; Weber, Ethofer and Ehlis, 2020 ). As was recently demonstrated (Kadosh and Staunton, 2019 ), psychological variables such as motivation or mood correlate with the success of neurofeedback. When feedback is presented in a continuous manner, it requires updating the feedback signal after each fMRI time point. Considering a TR for functional imaging of 2 seconds, we achieve a temporal resolution of 0.5 Hz in the feedback signal. This rather slow update rate of the feedback interface, coupled with the hemodynamic delay between the self-regulation and the feedback, may reduce the efficacy of the protocols. While decreasing the TR does not influence the rate at which neural processes occur nor the response time of the HRF, it may positively impact the immersiveness and the feedback calculation estimates for real-time use. Specifically, this increased temporal resolution could play a role in improving neurofeedback success, as it may contribute to the enhanced reliability of feedback signals based on activation or connectivity measures, providing a more realistic and immersive experience. In this study, we investigated our working hypothesis of interhemispheric functional connectivity in hMT + during a bistable visual motion task at four temporal resolutions: 2, 1.33, 1, and 0.4 Hz (TR = 0.5, 0.75, 1, and 2.5 seconds, respectively). This led to confirmation that one can link interhemispheric functional connectivity to global coherent perception. Then, we test if the activation metrics of hMT + are replicable across TRs. Finally, by establishing that perception can be read out from connectiviy patterns, we compare and contrast the characteristics of the feedback signal to be provided in a neurofeedback loop based on the activation and connectivity of hMT+. This region has been used in past studies as the target for neurofeedback, aiming for the self-modulation of the activity of this region in real time (Banca et al., 2015 ; Sousa et al., 2016 ; Direito et al., 2021 ). If we aimed to provide feedback on the perceptual switches based on the hMT + inter-hemispheric correlation, we checked how reliable and responsive this feedback would be at different resolutions. Methods Participants Fifteen healthy participants were recruited for this experiment (mean age 29.7 ± 8.4 years, 6 females) with normal or corrected-to-normal vision and no history of neurological or psychiatric diseases. All participants were right-handed, as confirmed by a handedness questionnaire adapted from (Oldfield, 1971 ): mean laterality index of 85.0 ± 9.0. All gave informed written consent before participating in accordance with the declaration of Helsinki, and the study followed the safety guidelines for magnetic resonance imaging research on humans. The work was approved by the Ethics Committee of the Faculty of Medicine of the University of Coimbra. fMRI data acquisition Scanning was performed on a 3T Siemens Magnetom Prisma fit, using a 64-channel head/neck coil, at the Institute of Nuclear Sciences Applied to Health (ICNAS), University of Coimbra, Portugal. The scanning session started with the acquisition of one 3D anatomical magnetization-prepared rapid acquisition gradient echo (MPRAGE) pulse sequence (TR = 2530 ms, echo time (TE) = 3.42 ms, flip angle (FA) = 7°, 176 slices, voxel size 1.0 × 1.0 × 1.0 mm 3 , field of view (FOV) = 256 × 256 mm 2 ). The functional runs were acquired using a 2D multi-band (MB) gradient-echo (GE) echo-planar imaging (EPI) sequence from the Center for Magnetic Resonance Research, University of Minnesota (Release R016a). We tested four different temporal resolutions: TR = 0.5 s (MB factor = 6, FA = 53º, 42 slices), TR = 0.75 s (MB factor = 4, FA = 63º, 40 slices), TR = 1 s (MB factor = 3, FA = 68º, 42 slices), TR = 2.5 s (MB factor = 1, FA = 85º, 42 slices). The remaining parameters were matched: TE = 30.2 ms, interleaved slices with 0.5 mm gap, voxel size 2.5 × 2.5 × 2.5 mm 3 , FoV 192 × 192 mm 2 (Table 1 ). Table 1 Main parameters of the four fMRI sequences. Sequence 1 2 3 4 TR (s) 0.5 0.75 1 2.5 TE (ms) 30.2 MB factor 6 4 3 1 Voxel size (mm) 2.5 Slice gap (mm) 0.5 Number of slices 42 40 42 42 Flip Angle (º) 53 63 68 85 Bandwidth (Hz/px) 2742 2742 2632 1994 Echo Spacing (ms) 0.51 0.49 0.49 0.59 Excite pulse duration (us) 2560 2560 2560 4840 EPI factor 76 GRAPPA Off Off Off 2 LeakBlock On On On Off For mapping and correction of image distortions related to magnetic field inhomogeneities, we acquired a pair of spin-echo images with anterior-posterior (AP) and posterior-anterior (PA) phase encoding polarity with matching geometry and echo-spacing to each of the functional scans. These were acquired before the functional runs. The participants’ physiological signals (respiration and pulse) were recorded during the functional runs using the scanner’s Physiological Measurement Unit (PMU). The respiratory signal was recorded at 50 Hz using a respiratory cushion, and the cardiac cycle was recorded at 200 Hz using a pulse sensor. Functional tasks We implemented three functional tasks based on a moving plaid stimulus. These stimuli are created by superimposing two gratings leading to a bistable percept: stimuli can be perceived moving coherently as a single surface (integration) or incoherently as two separate surfaces sliding over each other (segregation) (see supplementary materials and Fig. 1 ). The first is a localizer for the region of interest - hMT+. For this purpose, we created three conditions: i. ‘Fixation’ - a fixation cross; ii. ‘Static plaid’ - a stationary plaid; iii. ‘moving plaid’ - a moving plaid, which is inherently a bistable stimulus (coherent versus incoherent, e.g. integration versus segregation). The run lasted for 2.9 min and was composed of nine trials with the sequential presentation of each condition for 6 seconds. We named the following two tasks as ‘ambiguous’ and ‘unambiguous’ runs. These runs are composed of trials considering three conditions: ‘static’ (static plaid), ‘motion’ (ambiguous or unambiguous moving plaid), and ‘MAE’ (a period during which motion aftereffect is expected). During the ambiguous runs, the ‘motion’ condition showed the participants the moving plaid without any overlaid dots. As such, the stimulus is entirely ambiguous (the percept alternates between coherent and incoherent). Here, the participants were instructed to report the perceived type of motion (coherent or incoherent) using two buttons of a response box. During the unambiguous runs, the plaid is shown with overlaid dots moving either coherently down or incoherently inwards, which disambiguates the perception of the plaid (unambiguously coherent or incoherent). Based on the responses given by the participant in the previous ambiguous run, we manipulated the switches between coherent and incoherent motion in the unambiguous runs to match the previous responses precisely. This matched the time of each percept across both types of runs. The participants received the same instruction - to report the perceived type of motion at all times. In this work, we considered the localizer and the four unambiguous runs, one for each temporal resolution. fMRI data preprocessing The data were organized according to the Brain Imaging Data Structure (BIDS) (Gorgolewski et al., 2016 ), using BIDSkit (Tyszka, 2023 ) and dcm2niix (Li et al., 2016 ). Results included in this manuscript were obtained after data preprocessing performed using fMRIPrep 23.0.2 (Esteban et al., 2023 ), which is based on Nipype 1.8.6 (Gorgolewski et al., 2011 ; Esteban et al., 2022 ). For a detailed description of the fmriPrep pipeline please see supplementary materials. To perform quality checks on the anatomical and functional images and extract metrics for further analysis and sequence comparison we ran mriqc 23.0.1 (Esteban et al., 2017 ). All subsequent analyses were performed in Python using Nilearn (Abraham et al., 2014 ). fMRI data analysis Regarding the localizer data, hMT + was functionally localized for each subject using a standard GLM analysis. The design matrix included predictors for all experimental conditions (‘fixation’, ‘static plaid', ‘moving plaid’) and confound regressors based on the mean and first derivative of the voxels in the CSF and the six head motion parameters and their derivatives. Temporal high-pass filtering (cut-off = 0.03 Hz) and spatial smoothing (FWHM = 6 mm) were applied. hMT + was manually selected on both hemispheres based on the activation map for the contrast ‘moving plaid > static plaid’ corrected with Bonferroni p = 0.05, cluster threshold k = 50, and confirmed using the hMT + mask of Neurosynth (Yarkoni et al., 2011 )). Then, a spherical mask was designed around the center coordinates of each defined ROI (radius = 6 mm). The first step to analyzing the unambiguous runs was the definition of the coherent and incoherent predictors. To this end, using the perceptual reports acquired via button presses with a temporal resolution of 60 Hz, we calculated the percentage of coherent reports over the total number of reports for each volume. We used this value to label each volume as coherent or incoherent. This information defines which of the two percepts the participant is experiencing at any given time, allowing us to design predictors for coherent and incoherent events (for the GLM analyses) and to have the onsets and offsets of each percept to use for the neurofeedback algorithm. We estimated the functional activation maps of the unambiguous runs for each TR using a GLM. The design matrix included predictors for all experimental conditions (‘static’, ‘unambiguous motion’, ‘MAE’) and confound regressors based on the mean and first derivative of the voxels in the CSF and the six head motion parameters and their derivatives. Temporal high-pass filtering (cut-off = 0.003 Hz) was applied before extracting ROI activation measures (beta statistic and t-value). Our contrast of interest here was ‘unambiguous motion’ vs. ‘static’. We then used a different design matrix, replacing the ‘unambiguous motion’ predictor with predictors for coherent and incoherent percepts. As in the previous model, we extract the ROI beta statistic and t-value for the contrasts between coherent/incoherent and static plaid. For the following analyses, we extract the time course of the ROI and normalize it as a percent signal change to the mean value across time. We estimate the feedback based on this time course as we would in an actual neurofeedback experiment, by displaying the signal variation during the ‘unambiguous motion’ upregulation condition vs. the ‘static’ baseline/downregulation condition. Next, we use the time courses of the left and right hMT + to estimate Pearson’s correlation over time, a measure of inter-hemispheric functional connectivity. We use a sliding window of 6 seconds for the three lower TRs and of 7.5 seconds for the higher TR (a minimum of 3 data points was considered for obtaining a measurable correlation value). Using the perceptual report information, we center all correlation windows at the transitions between coherent and incoherent percepts and average them across trials and participants. Given the size of the windows, we only considered coherent and incoherent events that lasted at least 7.5 seconds. To use the functional connectivity information as feedback, we study its characteristics by defining a feedback rule. We defined that successful automatic detection of a perceptual transition meant a 10% minimum increase or decrease in correlation, for coherent and incoherent trials, respectively. We chose this threshold based on (Sousa et al., 2019 ), which reported an average difference in correlation between percepts of 0.13 ± 0.04. Based on this rule, we extract the transitions’ detection ratio and the time to detection from the first correlation window that includes the volume of transition. Data and code availability All the code for the above-mentioned analyses and data (including the activation maps and hMT+ timecourses) can be found at https://github.com/alexsayal/vpmb-tr. Results Localizer for hMT+ We defined the left and right hMT + ROIs in the localizer map of each individual, considering the contrast ‘moving’ vs. ‘static’, a voxel-wise correction with Bonferroni’s method (p = 0.05) and cluster-wise with a minimum cluster size of 50 voxels (Fig. 2 ). The MNI coordinates of the center of the left and right hMT + clusters for each subject are reported in Table S1 . The group activation map is shown in Figure S1 , and the average MNI coordinates of the center of the left and right clusters are (-42,-72,4) and (48,-66,6), respectively. GLM statistical values of hMT+ In Fig. 3 , we display the outcomes of the GLM analyses performed. In the first column, we plot the beta and t-values of the contrast ‘motion vs. static’. While there is no difference in the beta values between TRs, the t-value of the sequence with the lowest TR is significantly higher than the 2.5 s sequence. This indicates a higher level of certainty in the activation measurement (lower beta error) for the sequence with TR = 0.5 s. With the coherent and incoherent contrasts, we aimed to test if the measures of activation of rapidly changing events were higher and more reliable for the higher-resolution data. We found a significant difference in the t-value of the coherent contrast, again between the sequences with TR of 0.5 and 2.5 seconds. Interhemispheric connectivity In Fig. 4 , we display the time courses of interhemispheric correlation of hMT + for all the sequences with different temporal resolutions, centered on the perceptual switches between coherent and incoherent states and vice versa. The results are consistent across TRs - the transition to the coherent percept is linked to an increase in interhemispheric connectivity, while the transition to the incoherent percept is linked to a decrease in connectivity. BOLD feedback estimation based on the hMT + response We display the group average hMT + time course for each TR in Fig. 5 . This allows the visual comparison between the signals - the difference in temporal resolution is clear while the overall signal variation is similar across TRs. Using these time courses, we simulate a feedback signal to be provided during the motion/upregulation condition. We first normalize the signal to the mean value of the ‘static’, baseline condition, as a percent signal variation. Then, based on a maximum percent signal change of 2.5, we discretize this signal into 10 values, as we would if this information were to be displayed in a visual feedback interface with 10 activation levels. We display both of these signals in Fig. 6 for the four TRs. While all temporal resolutions provide a valid feedback signal, the variability for the 2.5 s sequence is higher (average SD = 4.97, while for the other sequences, it ranges from 1.39 to 1.60), with the feedback signal achieving values of zero during the upregulation condition. All signals take approximately the same time to reach the peak value of activation (as indicated by the red vertical dashed line). Feedback characteristics based on interhemispheric functional connectivity Here, we explored two characteristics of the feedback signal based on interhemispheric hMT + functional connectivity at different resolutions. Based on a rule that could distinguish coherent or incoherent perception, we calculated the percentage of detected switches (Fig. 7 ) and the time it took to detect such switches (Fig. 8 ) since the volume of transition for each sequence. The results show statistically significant differences in both metrics between the sequence with TR of 2.5 seconds and the other three sequences. Discussion Here we found that changes in interhemispheric connectivity reliably reflects perceptual integration and decision and becomes more evident with improved temporal resolution as implemented by multiband sequences. We also showed the feasibility of using connectivity based neurofeedback approaches Improving temporal resolution as a prior to study connectivity Here we used moving plaid stimuli, that lead to bistable percepts of coherent (reflecting integration) or non coherent (reflecting segregation) motion which lead to different activity levels in visual area hMT+, which may relate to differential profiles of adaptation and inhibition strengths across time for the two percepts (Sousa et al., 2018 , 2024 ; Sayal et al., 2020 ). Our first aim was to compare activation metrics across temporal resolutions for each of the conditions, coherent and incoherent motion percepts. The hypothesis was that higher temporal resolution sequences would allow for better activation estimates (i.e. smaller beta error values), which may be critical for subsequent connectivity analysis. Indeed, we found a significant difference between the t-values of the lower and higher temporal resolution sequences for the overall motion and coherent predictors, with higher t-values for the lower TR. This result indicates that higher resolution sequences might benefit rapid-changing events or short blocks, as they provide more points to estimate the GLM and improve the statistical outcomes. These results might, however, be specific to the research question, as (Darányi et al., 2021 ) found that increasing the temporal resolution of the signal did not universally improve the group-level statistical outcomes, reinforcing the importance of pilot testing and hemodynamic response function estimation for the ROIs being studied. Functional interhemispheric connectivity reliably reflects perceptual coherence A recent study emphasized the crucial role of replication by demonstrating that a meta-analytical approach, pooling information across teams, was essential in establishing a significant consensus on activated regions (Rotem Botvinik-Nezer et al. , 2020). We previously suggested using roof like stimuli at 7 tesla that interhemispheric functional connectivity reflects perceptual integration (Sousa et al., 2019 ). Remarkably this hypothesis lacks generalization. Here we confirmed this hypothesis using a different bistable stimulus at a lower field strength and distinct temporal resolutions. The results of (Sousa et al., 2019 ) showed opposing variations in hMT + inter-hemispheric correlation when changing either to coherent or incoherent percepts of an ambiguous visual stimulus. The connectivity between the left and right hMT + was shown to be critical in interpreting visual moving stimuli. When studying the “motion quartet”, an ambiguous stimulus that generates horizontal or vertical apparent motion, it was suggested that its perception requires interhemispheric integration, although functional connectivity was not studied (Genç et al., 2011 ). The authors were able to predict the perception based on structural connectivity data of callosal segments linking both hMT + regions. Howver these authors did not present functional connectivity data. Nevertheles, when the connectivity is manipulated using transcranial magnetic stimulation, the sensitivity to one of the possible percepts increases (Chiappini et al., 2022 ). When we analyze the volumes around the transition between coherent and incoherent perceptual states, as reported by the participants, we found a decrease in inter-hemispheric correlation after switching to the incoherent percept and an increase after switching to the coherent percept (Fig. 4 ), which is consistent with the findings of (Sousa et al., 2019 ) (see Fig. 4 of that study). One should note that this pioneering study uses a different ambiguous stimulus and a scanner with a higher static magnetic field (7T), but reaches consistent conclusions regarding this visual region activation when processing ambiguous moving stimuli. Moreover, in this study, the findings are replicated for all TRs. Functional connectivity as feedback of perceptual transitions We estimate the impact of temporal resolution on a BOLD-based feedback signal in Fig. 6 . An intriguing finding is that, for the sequence with the lowest resolution, the variability of the feedback signal is much higher. The standard deviation indicates that the feedback signal is less consistent across participants and that the mean even reaches zero at some points along the ‘upregulation’ block. This was not the case for the sequences with TR of 0.5 and 0.75 seconds. We found no difference in the time it takes for the signal from each sequence to reach the peak value of activation. Having established that the modulation of a single brain region in neurofeedback studies has an impact not only in the target region but also on a network of regions, neuroimaging studies move towards studying the connectivity between nodes of these networks (Ramot and Martin, 2022 ). In this sense, if one aims to use neurofeedback to target a specific function or behavior, one should characterize the functional network involved and the appropriate target region, assess if the changes we may be inducing are propagating along the rest of the network, or even if we should be targeting some parameter of the network itself. Here, we explore the interhemispheric functional connectivity of a region that responds to our stimuli, assessing its potential as a target for neuromodulation at different TRs. This has been successfully applied before in other neurofeedback paradigms (Pereira et al., 2019 ; Wang et al., 2020 ; Hayashi et al., 2022 ). Specifically, we simulated the feedback of the perceptual switches based on the hMT + interhemispheric correlation value. We found that the feedback information is both more reliable and more responsive for the higher resolution sequences (Figs. 7 and 8 ) when we consider a simple rule for determining a perceptual switch. Based on a previous study, which found that the more information available to the participant during neurofeedback, the greater its ability to self-regulate brain activity (Sorger et al., 2018 ), we conclude that the time delay between the modulation and the feedback of that modulation time is an important factor in neurofeedback paradigms. As such, if the goal is to provide feedback based on connectivity, one could consider using higher temporal resolution sequences. Trade-offs, limitations and compromises The neural processes indirectly inferred using fMRI vary in the scale of the millisecond but are sampled in the scale of the second. This represents a severe downsampling with fMRI sample intervals typically ranging from 1 to 3 s, way longer than typical interneuron delays (Seth, Chorley and Barnett, 2013 ). Nevertheless, this measure of neuronal activity can be interpreted as a representation of a low pass filter of the local field potentials, showing that neural patterns can also be found in the BOLD signal, a particularly relevant feature for connectivity methods applied to fMRI data (Fernandes et al., 2020 ). While the temporal resolution of this neuroimaging technique is far from others, such as electroencephalography, lowering the TR can provide us with a better sampling of the HRF and more certainty when studying brain connectivity with functional MRI. Increasing the temporal resolution during real-time acquisitions poses challenges both in image quality and technical implementations. As recent research suggests, multiband sequences may diminish the detection ability of activity in mesolimbic regions (Srirangarajan et al., 2021 ) but not in other regions in the brain, as we have seen here with hMT+. Also, the temporal SNR is not homogeneous in the whole brain and tends to be lower with higher multiband factors (Todd et al., 2016 ; Srirangarajan et al., 2021 ). Hence, the recommendation for the use of these sequences for real-time fMRI is highly dependent on the target region or network of interest. From the technical point of view, fMRI neurofeedback has high computational requirements, namely regarding image reconstruction at the scanner, network capacity and speed for image transfer at every volume, and the real-time image processing software itself. Some of these requirements have been addressed in (Renz et al., 2023 ), where the authors managed to achieve a stable real-time acquisition with a TR of 1 second, but not less, a result which may be highly dependent on each site’s MRI setup and infrastructure. Conclusion In conclusion, our study generalizes the hypothesis that interhemispheric connectivity reflects spontaneous perceptual switches in perceptual coherence, which is shown across temporal resolutions (even at 3T). We found stable activation metrics in the hMT + region across resolutions and consistent results of changes in interhemispheric connectivity as a function of perceptual decision, affirming the reliability of fMRI for studying dynamic connectivity patterns. While simulating a feedback signal based on interhemispheric connectivity, higher resolution sequences (TRs within 1 second and lower) showed benefits. Overall, integrating fast fMRI sequences may offer promising avenues for optimizing cognitive functions within neurofeedback settings. Declarations Data and code availability All the code for the above-mentioned analyses and data (including the activation maps and hMT+ timecourses) can be found at https://github.com/alexsayal/vpmb-tr. Author contributions – AS, TS and MCB designed the study ,BD, TS and MCB provided supervision, AS acquired and analysed the data, SA acquired the data, AS wrote the first draft of the manuscript, all authors revised, edited and approved the final version. Competing interests – the authors report no competing interests Funding : This research work was funded by the Portuguese Foundation for Science and Technology (FCT) (grants: UID/04950B/2020, UID/04950P/2020, DSAIPA/DS/0041/2020, PTDC/PSI-GER/1326/2020, 2022.02963.PTDC – Hallucin and by the BIAL Foundation project 207/16. References Abraham, A., Pedregosa, F., Eickenberg, M., Gervais, P., Mueller, A., Kossaifi, J., Gramfort, A., Thirion, B. and Varoquaux, G. (2014) ‘Machine learning for neuroimaging with scikit-learn’, Frontiers in Neuroinformatics , 8. Available at: https://doi.org/10.3389/fninf.2014.00014. Banca, P., Sousa, T., Catarina Duarte, I. and Castelo-Branco, M. (2015) ‘Visual motion imagery neurofeedback based on the hMT+/V5 complex: evidence for a feedback-specific neural circuit involving neocortical and cerebellar regions’, Journal of Neural Engineering , 12(6), p. 066003. Available at: https://doi.org/10.1088/1741-2560/12/6/066003. Bhandari, R., Kirilina, E., Caan, M., Suttrup, J., De Sanctis, T., De Angelis, L., Keysers, C. and Gazzola, V. (2020) ‘Does higher sampling rate (multiband + SENSE) improve group statistics - An example from social neuroscience block design at 3T’, NeuroImage , 213, p. 116731. Available at: https://doi.org/10.1016/j.neuroimage.2020.116731. Chen, L., T. Vu, A., Xu, J., Moeller, S., Ugurbil, K., Yacoub, E. and Feinberg, D.A. (2015) ‘Evaluation of highly accelerated simultaneous multi-slice EPI for fMRI’, NeuroImage , 104, pp. 452–459. Available at: https://doi.org/10.1016/j.neuroimage.2014.10.027. Chiappini, E., Sel, A., Hibbard, P.B., Avenanti, A. and Romei, V. (2022) ‘Increasing interhemispheric connectivity between human visual motion areas uncovers asymmetric sensitivity to horizontal motion’, Current Biology , 32(18), pp. 4064-4070.e3. Available at: https://doi.org/10.1016/j.cub.2022.07.050. Darányi, V., Hermann, P., Homolya, I., Vidnyánszky, Z. and Nagy, Z. (2021) ‘An empirical investigation of the benefit of increasing the temporal resolution of task-evoked fMRI data with multi-band imaging’, Magnetic Resonance Materials in Physics, Biology and Medicine , 34(5), pp. 667–676. Available at: https://doi.org/10.1007/s10334-021-00918-z. Demetriou, L., Kowalczyk, O.S., Tyson, G., Bello, T., Newbould, R.D. and Wall, M.B. (2018) ‘A comprehensive evaluation of increasing temporal resolution with multiband-accelerated protocols and effects on statistical outcome measures in fMRI’, NeuroImage , 176, pp. 404–416. Available at: https://doi.org/10.1016/j.neuroimage.2018.05.011. Direito, B., Ramos, M., Pereira, J., Sayal, A., Sousa, T. and Castelo-Branco, M. (2021) ‘Directly Exploring the Neural Correlates of Feedback-Related Reward Saliency and Valence During Real-Time fMRI-Based Neurofeedback’, Frontiers in Human Neuroscience , 14, p. 578119. Available at: https://doi.org/10.3389/fnhum.2020.578119. Duarte, J.V., Costa, G.N., Martins, R. and Castelo‐Branco, M. (2017) ‘Pivotal role of hMT+ in long‐range disambiguation of interhemispheric bistable surface motion’, Human Brain Mapping , 38(10), pp. 4882–4897. Available at: https://doi.org/10.1002/hbm.23701. Esteban, O., Birman, D., Schaer, M., Koyejo, O.O., Poldrack, R.A. and Gorgolewski, K.J. (2017) ‘MRIQC: Advancing the automatic prediction of image quality in MRI from unseen sites’, PLOS ONE . Edited by B.C. Bernhardt, 12(9), p. e0184661. Available at: https://doi.org/10.1371/journal.pone.0184661. Esteban, O., Markiewicz, C.J., Burns, C., Goncalves, M., Jarecka, D., Ziegler, E., Berleant, S., Ellis, D.G., Pinsard, B., Madison, C., Waskom, M., Notter, M.P., Clark, Daniel, Manhães-Savio, A., Clark, Dav, Jordan, K., Dayan, M., Halchenko, Y.O., Loney, F., Norgaard, M., Salo, T., Dewey, B.E., Johnson, H., Bougacha, S., Keshavan, A., Yvernault, B., Hamalainen, C., Christian, H., Ćirić, R., Dubois, M., Joseph, M., Cipollini, B., Tilley II, S., Visconti di Oleggio Castello, M., De La Vega, A., Wong, J., Kaczmarzyk, J., Huntenburg, J.M., Clark, M.G., Kent, J.D., Benderoff, E., Erickson, D., Dias, M. de F., Hanke, M., Giavasis, S., Moloney, B., Nichols, B.N., Tungaraza, R., Dell’Orco, A., Frohlich, C., Wassermann, D., de Hollander, G., Koudoro, S., Eshaghi, A., Millman, J., Mancini, M., Close, T., Nielson, D.M., Varoquaux, G., Waller, L., Watanabe, A., Mordom, D., Guillon, J., Robert-Fitzgerald, T., Chetverikov, A., Rokem, A., Acland, B., Forbes, J., Markello, R., Gillman, A., Bernardoni, F., Kong, X.-Z., Geisler, D., Salvatore, J., Gramfort, A., Doll, A., Buchanan, C., DuPre, E., Liu, S., Schaefer, A., Kleesiek, J., Sikka, S., Schwartz, Y., Ghayoor, A., Lee, J.A., Mattfeld, A., Richie-Halford, A., Liem, F., Vaillant, G., Perez-Guevara, M.F., Heinsfeld, A.S., Haselgrove, C., Durnez, J., Lampe, L., Poldrack, R., Glatard, T., Baratz, Z., Tabas, A., Cumba, C., Pérez-García, F., Blair, R., Iqbal, S., Welch, D., Ben-Zvi, G., Contier, O., Triplett, W., Craddock, R.C., Correa, C., Papadopoulos Orfanos, D., Stadler, J., Warner, J., Sisk, L.M., Falkiewicz, M., Sharp, P., Rothmei, S., Kim, S., Weinstein, A., Kahn, A.E., Kastman, E., Bottenhorn, K., Grignard, M., Perkins, L.N., Zhou, D., Bielievtsov, D., Cooper, G., Stojic, H., Hui Qian, T., Linkersdörfer, J., Renfro, M., Hinds, O., Stanley, O., Küttner, R., Pauli, W.M., Xie, X., Glen, D., Kimbler, A., Meyers, B., Tarbert, C., Ginsburg, D., Haehn, D., Margulies, D.S., Condamine, E., Ma, F., Malone, I.B., Snoek, L., Brett, M., Cieslak, M., Hallquist, M., Molina-Romero, M., Bilgel, M., Lee, N., Kuntke, P., Jalan, R., Inati, S., Gerhard, S., Mathotaarachchi, S., Saase, V., Van, A., Steele, C.J., Ort, E., Lerma-Usabiaga, G., Schwabacher, I., Arias, J., Lai, J., Pellman, J., Huguet, J., WEN, J., Leinweber, K., Chawla, K., Weninger, L., Modat, M., Mukhometzianov, R., Harms, R., Andberg, S.K., Matsubara, K., González Orozco, A.A., Routier, A., Marina, A., Davison, A., Floren, A., Park, A., Frederick, B., Cheung, B., McDermottroe, C., McNamee, D., Shachnev, D., Vogel, D., Flandin, G., Jones, H., Gonzalez, I., Varada, J., Schlamp, K., Podranski, K., Huang, L., Noel, M., Pannetier, N., Numssen, O., Khanuja, R., Urchs, S., Shim, S., Nickson, T., Huang, L., Broderick, W., Tambini, A., Mihai, P.G., Gorgolewski, K.J. and Ghosh, S. (2022) ‘nipy/nipype: 1.8.3’. Zenodo. Available at: https://doi.org/10.5281/ZENODO.596855. Esteban, O., Markiewicz, C.J., Goncalves, M., Provins, C., Kent, J.D., DuPre, E., Salo, T., Ciric, R., Pinsard, B., Blair, R.W., Poldrack, R.A. and Gorgolewski, K.J. (2023) ‘fMRIPrep: a robust preprocessing pipeline for functional MRI’. Zenodo. Available at: https://doi.org/10.5281/ZENODO.852659. Fernandes, T.T., Direito, B., Sayal, A., Pereira, J., Andrade, A. and Castelo-Branco, M. (2020) ‘The boundaries of state-space Granger causality analysis applied to BOLD simulated data: A comparative modelling and simulation approach’, Journal of Neuroscience Methods , 341, p. 108758. Available at: https://doi.org/10.1016/j.jneumeth.2020.108758. Genç, E., Bergmann, J., Singer, W. and Kohler, A. (2011) ‘Interhemispheric Connections Shape Subjective Experience of Bistable Motion’, Current Biology , 21(17), pp. 1494–1499. Available at: https://doi.org/10.1016/j.cub.2011.08.003. Gorgolewski, K., Burns, C.D., Madison, C., Clark, D., Halchenko, Y.O., Waskom, M.L. and Ghosh, S.S. (2011) ‘Nipype: A Flexible, Lightweight and Extensible Neuroimaging Data Processing Framework in Python’, Frontiers in Neuroinformatics , 5. Available at: https://doi.org/10.3389/fninf.2011.00013. Gorgolewski, K.J., Auer, T., Calhoun, V.D., Craddock, R.C., Das, S., Duff, E.P., Flandin, G., Ghosh, S.S., Glatard, T., Halchenko, Y.O., Handwerker, D.A., Hanke, M., Keator, D., Li, X., Michael, Z., Maumet, C., Nichols, B.N., Nichols, T.E., Pellman, J., Poline, J.-B., Rokem, A., Schaefer, G., Sochat, V., Triplett, W., Turner, J.A., Varoquaux, G. and Poldrack, R.A. (2016) ‘The brain imaging data structure, a format for organizing and describing outputs of neuroimaging experiments’, Scientific Data , 3(1), p. 160044. Available at: https://doi.org/10.1038/sdata.2016.44. Hampson, M., Ruiz, S. and Ushiba, J. (2020) ‘Neurofeedback’, NeuroImage , 218, p. 116473. Available at: https://doi.org/10.1016/j.neuroimage.2019.116473. Hayashi, M., Okuyama, K., Mizuguchi, N., Hirose, R., Okamoto, T., Kawakami, M. and Ushiba, J. (2022) ‘Spatially bivariate EEG-neurofeedback can manipulate interhemispheric inhibition’, eLife , 11, p. e76411. Available at: https://doi.org/10.7554/eLife.76411. Jahanian, H., Holdsworth, S., Christen, T., Wu, H., Zhu, K., Kerr, A.B., Middione, M.J., Dougherty, R.F., Moseley, M. and Zaharchuk, G. (2019) ‘Advantages of short repetition time resting-state functional MRI enabled by simultaneous multi-slice imaging’, Journal of Neuroscience Methods , 311, pp. 122–132. Available at: https://doi.org/10.1016/j.jneumeth.2018.09.033. Kadosh, K.C. and Staunton, G. (2019) ‘A systematic review of the psychological factors that influence neurofeedback learning outcomes’, NeuroImage , 185, pp. 545–555. Available at: https://doi.org/10.1016/j.neuroimage.2018.10.021. Li, X., Morgan, P.S., Ashburner, J., Smith, J. and Rorden, C. (2016) ‘The first step for neuroimaging data analysis: DICOM to NIfTI conversion’, Journal of Neuroscience Methods , 264, pp. 47–56. Available at: https://doi.org/10.1016/j.jneumeth.2016.03.001. Oldfield, R.C. (1971) ‘The assessment and analysis of handedness: The Edinburgh inventory’, Neuropsychologia , 9(1), pp. 97–113. Available at: https://doi.org/10.1016/0028-3932(71)90067-4. Pereira, J., Direito, B., Sayal, A., Ferreira, C. and Castelo-Branco, M. (2019) ‘Self-modulation of premotor cortex interhemispheric connectivity in a real-time functional magnetic resonance imaging neurofeedback study using an adaptive approach’, Brain Connectivity , 9(9), pp. 662–672. Ramot, M. and Martin, A. (2022) ‘Closed-loop neuromodulation for studying spontaneous activity and causality’, Trends in Cognitive Sciences , 26(4), pp. 290–299. Available at: https://doi.org/10.1016/j.tics.2022.01.008. Renz, M.P., Zidda, F., Andoh, J., Prager, M., Sack, M., Becker, R., Ruf, M., Schmitgen, M.M., Wolf, R.C., Meyer‐Lindenberg, A. and Tost, H. (2023) ‘Practical challenges of continuous real‐time functional magnetic resonance imaging neurofeedback with multiband accelerated echo‐planar imaging and short repetition times’, Human Brain Mapping , 44(3), pp. 1278–1282. Available at: https://doi.org/10.1002/hbm.26154. Risk, B.B., Kociuba, M.C. and Rowe, D.B. (2018) ‘Impacts of simultaneous multislice acquisition on sensitivity and specificity in fMRI’, NeuroImage , 172, pp. 538–553. Available at: https://doi.org/10.1016/j.neuroimage.2018.01.078. Ros, T., Enriquez-Geppert, S., Zotev, V., Young, K.D., Wood, G., Whitfield-Gabrieli, S., Wan, F., Vuilleumier, P., Vialatte, F., Van De Ville, D., Todder, D., Surmeli, T., Sulzer, J.S., Strehl, U., Sterman, M.B., Steiner, N.J., Sorger, B., Soekadar, S.R., Sitaram, R., Sherlin, L.H., Schönenberg, M., Scharnowski, F., Schabus, M., Rubia, K., Rosa, A., Reiner, M., Pineda, J.A., Paret, C., Ossadtchi, A., Nicholson, A.A., Nan, W., Minguez, J., Micoulaud-Franchi, J.-A., Mehler, D.M.A., Lührs, M., Lubar, J., Lotte, F., Linden, D.E.J., Lewis-Peacock, J.A., Lebedev, M.A., Lanius, R.A., Kübler, A., Kranczioch, C., Koush, Y., Konicar, L., Kohl, S.H., Kober, S.E., Klados, M.A., Jeunet, C., Janssen, T.W.P., Huster, R.J., Hoedlmoser, K., Hirshberg, L.M., Heunis, S., Hendler, T., Hampson, M., Guggisberg, A.G., Guggenberger, R., Gruzelier, J.H., Göbel, R.W., Gninenko, N., Gharabaghi, A., Frewen, P., Fovet, T., Fernández, T., Escolano, C., Ehlis, A.-C., Drechsler, R., Christopher deCharms, R., Debener, S., De Ridder, D., Davelaar, E.J., Congedo, M., Cavazza, M., Breteler, M.H.M., Brandeis, D., Bodurka, J., Birbaumer, N., Bazanova, O.M., Barth, B., Bamidis, P.D., Auer, T., Arns, M. and Thibault, R.T. (2020) ‘Consensus on the reporting and experimental design of clinical and cognitive-behavioural neurofeedback studies (CRED-nf checklist)’, Brain , 143(6), pp. 1674–1685. Available at: https://doi.org/10.1093/brain/awaa009. Rotem Botvinik-Nezer, Felix Holzmeister, Lei Zhang, Xu Zhang, Joshua E. Zosky, Thomas E. Nichols, Russell A. Poldrack, and Tom Schonberg (2020) ‘Variability in the analysis of a single neuroimaging dataset by many teams’, Nature [Preprint]. Available at: https://doi.org/10.1038/s41586-020-2314-9. Sayal, A., Sousa, T., Duarte, J.V., Costa, G.N., Martins, R. and Castelo-Branco, M. (2020) ‘Identification of competing neural mechanisms underlying positive and negative perceptual hysteresis in the human visual system’, NeuroImage , 221, p. 117153. Scharnowski, F., Hutton, C., Josephs, O., Weiskopf, N. and Rees, G. (2012) ‘Improving Visual Perception through Neurofeedback’, The Journal of Neuroscience , 32(49), pp. 17830–17841. Available at: https://doi.org/10.1523/JNEUROSCI.6334-11.2012. Seth, A.K., Chorley, P. and Barnett, L.C. (2013) ‘Granger causality analysis of fMRI BOLD signals is invariant to hemodynamic convolution but not downsampling’, NeuroImage , 65, pp. 540–555. Available at: https://doi.org/10.1016/j.neuroimage.2012.09.049. Sitaram, R., Ros, T., Stoeckel, L., Haller, S., Scharnowski, F., Lewis-Peacock, J., Weiskopf, N., Blefari, M.L., Rana, M., Oblak, E., Birbaumer, N. and Sulzer, J. (2017) ‘Closed-loop brain training: the science of neurofeedback’, Nature Reviews Neuroscience , 18(2), pp. 86–100. Available at: https://doi.org/10.1038/nrn.2016.164. Sorger, B., Kamp, T., Weiskopf, N., Peters, J.C. and Goebel, R. (2018) ‘When the Brain Takes “BOLD” Steps: Real-Time fMRI Neurofeedback Can Further Enhance the Ability to Gradually Self-regulate Regional Brain Activation’, Neuroscience , 378, pp. 71–88. Available at: https://doi.org/10.1016/j.neuroscience.2016.09.026. Sousa, T., Direito, B., Lima, J., Ferreira, C., Nunes, U. and Castelo-Branco, M. (2016) ‘Control of Brain Activity in hMT+/V5 at Three Response Levels Using fMRI-Based Neurofeedback/BCI’, PLOS ONE . Edited by J. Bodurka, 11(5), p. e0155961. Available at: https://doi.org/10.1371/journal.pone.0155961. Sousa, T., Duarte, J.V., Costa, G.N., Kemper, V.G., Martins, R., Goebel, R. and Castelo-Branco, M. (2019) ‘Tracking perceptual decision mechanisms through changes in interhemispheric functional connectivity in human visual cortex’, Scientific Reports , 9(1), p. 1242. Available at: https://doi.org/10.1038/s41598-018-37822-x. Sousa, T., Sayal, A., Duarte, J.V., Costa, G.N. and Castelo-Branco, M. (2024) ‘A human cortical adaptive mutual inhibition circuit underlying competition for perceptual decision and repetition suppression reversal’, NeuroImage , 285, p. 120488. Available at: https://doi.org/10.1016/j.neuroimage.2023.120488. Sousa, T., Sayal, A., Duarte, J.V., Costa, G.N., Martins, R. and Castelo-Branco, M. (2018) ‘Evidence for distinct levels of neural adaptation to both coherent and incoherently moving visual surfaces in visual area hMT+’, NeuroImage , 179, pp. 540–547. Available at: https://doi.org/10.1016/j.neuroimage.2018.06.075. Srirangarajan, T., Mortazavi, L., Bortolini, T., Moll, J. and Knutson, B. (2021) ‘Multi‐band FMRI compromises detection of mesolimbic reward responses’, NeuroImage , 244, p. 118617. Available at: https://doi.org/10.1016/j.neuroimage.2021.118617. Todd, N., Josephs, O., Zeidman, P., Flandin, G., Moeller, S. and Weiskopf, N. (2017) ‘Functional Sensitivity of 2D Simultaneous Multi-Slice Echo-Planar Imaging: Effects of Acceleration on g-factor and Physiological Noise’, Frontiers in Neuroscience , 11. Available at: https://doi.org/10.3389/fnins.2017.00158. Todd, N., Moeller, S., Auerbach, E.J., Yacoub, E., Flandin, G. and Weiskopf, N. (2016) ‘Evaluation of 2D multiband EPI imaging for high-resolution, whole-brain, task-based fMRI studies at 3T: Sensitivity and slice leakage artifacts’, NeuroImage , 124, pp. 32–42. Available at: https://doi.org/10.1016/j.neuroimage.2015.08.056. Tyszka, M. (2023) ‘bidskit’. Available at: https://github.com/jmtyszka/bidskit. Wall, M.B. (2023) ‘Multiband acquisition sequences for fMRI: Proceed with caution’, Aperture Neuro , 3. Available at: https://doi.org/10.52294/001c.91292. Wang, T., Peeters, R., Mantini, D. and Gillebert, C.R. (2020) ‘Modulating the interhemispheric activity balance in the intraparietal sulcus using real-time fMRI neurofeedback: Development and proof-of-concept’, NeuroImage: Clinical , 28, p. 102513. Available at: https://doi.org/10.1016/j.nicl.2020.102513. Weber, L.A., Ethofer, T. and Ehlis, A.-C. (2020) ‘Predictors of neurofeedback training outcome: A systematic review’, NeuroImage: Clinical , 27, p. 102301. Available at: https://doi.org/10.1016/j.nicl.2020.102301. Yarkoni, T., Poldrack, R.A., Nichols, T.E., Van Essen, D.C. and Wager, T.D. (2011) ‘Large-scale automated synthesis of human functional neuroimaging data’, Nature Methods , 8(8), pp. 665–670. Available at: https://doi.org/10.1038/nmeth.1635. Zhang, Y., Ye, Z., Xia, C., Tan, Y., Zhang, M., Lv, X., Tang, J. and Li, Z. (2024) ‘Clinical Applications and Recent Updates of Simultaneous Multi-slice Technique in Accelerated MRI’, Academic Radiology , p. S1076633223007109. Available at: https://doi.org/10.1016/j.acra.2023.12.032. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4335511","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":298232743,"identity":"7d5a5857-7a78-4492-bae8-bb0472eb58ba","order_by":0,"name":"Alexandre Sayal","email":"","orcid":"","institution":"University of Coimbra","correspondingAuthor":false,"prefix":"","firstName":"Alexandre","middleName":"","lastName":"Sayal","suffix":""},{"id":298232744,"identity":"b1105542-5463-410f-823e-c10f64c95d1e","order_by":1,"name":"Bruno Direito","email":"","orcid":"","institution":"University of Coimbra","correspondingAuthor":false,"prefix":"","firstName":"Bruno","middleName":"","lastName":"Direito","suffix":""},{"id":298232745,"identity":"f96a8d4a-c90a-4b46-9fc7-45fe9642fe3e","order_by":2,"name":"Teresa Sousa","email":"","orcid":"","institution":"University of Coimbra","correspondingAuthor":false,"prefix":"","firstName":"Teresa","middleName":"","lastName":"Sousa","suffix":""},{"id":298232746,"identity":"0928d1ea-fc0e-4c81-afec-1b3d7dae8ce7","order_by":3,"name":"João Duarte","email":"","orcid":"","institution":"Champalimaud Foundation","correspondingAuthor":false,"prefix":"","firstName":"João","middleName":"","lastName":"Duarte","suffix":""},{"id":298232747,"identity":"b3489025-fcfa-4a67-88f3-52d14a0b544e","order_by":4,"name":"Sónia Afonso","email":"","orcid":"","institution":"University of Coimbra","correspondingAuthor":false,"prefix":"","firstName":"Sónia","middleName":"","lastName":"Afonso","suffix":""},{"id":298232748,"identity":"46e3b9b7-0ca0-4007-a6f1-00fb0c8c252a","order_by":5,"name":"Miguel Castelo-Branco","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAyUlEQVRIiWNgGAWjYFACHgjFz3wAziZSi2RbAozNTKQWg2PEatFtP3vsw88ddnLGx7gTH7xhOCzPIN1/AK8WszN5yTN7zyQbmx3j3Ww4h+GwYYPMYfy2mN3gMWbgbTuQuO1+7zZpHobDCQwSyYS1MP4Fatncxrv9N9FamEG2bGDj3cZMnBagX5hl25KNJYB+kZxjkG7YJnPYAL+W42cPM75ts5Pjb+Pd+OFNhbU8v3TjA/zWoAKg+WwSpGiAADK0jIJRMApGwfAGANYIP9jmCOGTAAAAAElFTkSuQmCC","orcid":"","institution":"University of Coimbra","correspondingAuthor":true,"prefix":"","firstName":"Miguel","middleName":"","lastName":"Castelo-Branco","suffix":""}],"badges":[],"createdAt":"2024-04-27 21:54:07","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4335511/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4335511/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":56121744,"identity":"d0739a79-6b63-4428-ba88-618c1e821267","added_by":"auto","created_at":"2024-05-08 20:29:17","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":49743,"visible":true,"origin":"","legend":"\u003cp\u003eStimuli used in the stage 1 experiment. A) Functional localizer used to map V1 and hMT+ in each participant's visual cortex. Moving dots were shown inside a circular aperture at the center of a black screen. Panels B, C, and D illustrate a plaid stimulus. By superimposing two gratings (B), moving orthogonally to the lines, a bistable stimulus is created, which can be perceived moving coherently (C) as a single surface or incoherently (D) as two separate surfaces sliding over each other. The arrows illustrate the direction of perceived motion. E) Plaid stimulus used during the unambiguous runs. Depending on the moving dots' direction, the otherwise ambiguous stimulus was readily perceived as a plaid moving coherently or incoherently. Adapted from (Sousa et al. 2018).\u003c/p\u003e","description":"","filename":"1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4335511/v1/18316e85bdd92fb1de9bbc9e.jpg"},{"id":56121742,"identity":"8b5d2eb8-fabe-4ad3-b4bc-cd98abc53572","added_by":"auto","created_at":"2024-05-08 20:29:17","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":84628,"visible":true,"origin":"","legend":"\u003cp\u003eGLM maps for the hMT+ localizer of each participant, contrasting the moving with the static plaid conditions (z values corrected voxel-wise with Bonferroni’s correction and threshold p = 0.05 and cluster-wise with a minimum cluster size of 50 voxels). The maps show z-values between the estimated threshold per subject and 20, over the axial slice with Z = 4.\u003c/p\u003e","description":"","filename":"2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4335511/v1/b6baf2faa754e9537558a9a6.jpg"},{"id":56121743,"identity":"362ff9cd-8f2c-454a-8997-0f400bec1a6b","added_by":"auto","created_at":"2024-05-08 20:29:17","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":87609,"visible":true,"origin":"","legend":"\u003cp\u003eComparing the beta and t-values for the activation of hMT+ during the functional runs for each TR and contrast of interest - ‘motion’, ‘coherent’, and ‘incoherent’ vs. ‘static’. We display boxplots (median, quartiles, standard deviation) with overlaid data points and the statistical pairwise comparison was performed using two-sided Mann-Whitney-Wilcoxon tests with Bonferroni’s correction (p=0.05). Each data point corresponds to one subject and only significant comparisons are drawn.\u003c/p\u003e","description":"","filename":"3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4335511/v1/9577462e65278a6b31f5629b.jpg"},{"id":56122553,"identity":"b9233f03-b4ed-4486-8402-15ceadf1350b","added_by":"auto","created_at":"2024-05-08 20:37:17","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":120531,"visible":true,"origin":"","legend":"\u003cp\u003eInter-hemispheric hMT+ correlation time courses (window size of 6 seconds / 7.5 seconds for TR=2.5s) per TR centered on coherent (top row) and incoherent (bottom row) perceptual switches. A dashed vertical line indicates the volume of transition. The first and last correlation points correspond to windows that do not include the volume of transition. Linear regression was performed on the data highlighted in green using ordinary least squares.\u003c/p\u003e","description":"","filename":"4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4335511/v1/5eca1e0ae8d49f2995bd321b.jpg"},{"id":56121749,"identity":"192fd604-6e6c-45c4-a9f5-ddceedf40a57","added_by":"auto","created_at":"2024-05-08 20:29:17","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":58261,"visible":true,"origin":"","legend":"\u003cp\u003eComparing the average hMT+ time course resolution for each TR. The motion (up-regulation) blocks are represented in light blue and the signals represent the percent BOLD signal variation to the mean.\u003c/p\u003e","description":"","filename":"5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4335511/v1/60a4789ff158efc053aa1215.jpg"},{"id":56121746,"identity":"0b917bc6-1ce0-4b35-b681-f7c9f6996944","added_by":"auto","created_at":"2024-05-08 20:29:17","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":164892,"visible":true,"origin":"","legend":"\u003cp\u003eLeft panels - Event-related averages for each TR. The timecourses were normalized to the mean of the static periods, considering a hemodynamic delay of 4 seconds. The red dashed line indicates the first point with at least 2.5% variation. Right panels - average of the simulated feedback signals (10 levels, max PSC = 2.5 %). The shadowed areas represent the standard error of the mean.\u003c/p\u003e","description":"","filename":"6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4335511/v1/ee1f1ea842142e48a2fa73c1.jpg"},{"id":56121747,"identity":"8e09cab7-b4c4-4bfc-83d6-dc93bc0cc971","added_by":"auto","created_at":"2024-05-08 20:29:17","extension":"jpg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":41156,"visible":true,"origin":"","legend":"\u003cp\u003ePercentage of identified perceptual switches based on the correlation signal over the total number of perceptual switches reported by the participant for each TR, based on the rule of 10% increase or decrease in correlation for coherent and incoherent trials, respectively. We display the results of a two-tailed Mann-Whitney-Wilcoxon test with Bonferroni correction (p = 0.05) on the coherent trials data in blue.\u003c/p\u003e","description":"","filename":"7.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4335511/v1/d99c037d3d7deb932f997d4f.jpg"},{"id":56122555,"identity":"744a11d8-53e9-4470-b2f5-dc159ec4e425","added_by":"auto","created_at":"2024-05-08 20:37:17","extension":"jpg","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":44540,"visible":true,"origin":"","legend":"\u003cp\u003eBoxplot representation of the time to detection. For each sequence, we estimate how much time it takes to provide feedback after the perceptual switch. A two-tailed Mann-Whitney-Wilcoxon test with Bonferroni correction (p = 0.05) was applied to the coherent and incoherent trial data.\u003c/p\u003e","description":"","filename":"8.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4335511/v1/5d80c032cdb329571dac1009.jpg"},{"id":65294781,"identity":"372a75ed-55e6-4005-a5f2-cc8f2f788bcb","added_by":"auto","created_at":"2024-09-25 18:46:40","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1130636,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4335511/v1/d9d62a57-528f-4d96-9e92-c077dec6acca.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Interhemispheric connectivity during spontaneous switches in visual perceptual coherence as revealed by fMRI at multiple temporal resolutions","fulltext":[{"header":"Introduction","content":"\u003cp\u003eFunctional magnetic resonance imaging (fMRI) has revolutionized our understanding of brain function, and it is now possible to observe neural activity in real-time. One critical aspect in the context of fMRI is its temporal resolution. Unlike other neuroimaging techniques, such as electroencephalography (EEG) or magnetoencephalography (MEG), fMRI has limited temporal resolution, meaning it may not have the sensitivity to capture the hemodynamic responses to rapid changes in neural activity. This aspect becomes particularly significant when considering the study of dynamic connectivity patterns in the human visual system. The perceptual integration of components of motion to generate a global pattern may require communication between different parts of the visual system (Duarte et al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). In the case of bistable perception, the fast changes in interhemispheric connectivity inform us about the binding mechanisms of visual motion features (Sousa et al., \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). While aiming for a novel intervention that target online perceptual learning mechanisms (Scharnowski et al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2012\u003c/span\u003e), the detection in real-time of functional connectivity changes is key. Hence, in this study, we explore the interhemispheric connectivity in the visual cortex in response to bistable (coherent or incoherent) moving stimuli at different temporal resolutions, to validate and infer perceptual states based on this metric. Online tracking of perception might also help prompt the exploration of a real-time fMRI connectivity based neurofeedback protocol that aims at potentiating cognitive functions in an immersive therapeutic setting. Ultimately, this output of neuroscientific relevance to track interhemispheric integration and perceptual decision would also allows us to generalize the impact of temporal resolution in real time approaches such as neurofeedback paradigms.\u003c/p\u003e \u003cp\u003eOur first goal is to test the hypothesis that the functional dynamics of the human motion complex (hMT+) across hemispheres reflects bistable global motion perception. This region, homologous to the monkey MT region, is a well-known region of the visual cortex that responds to motion in the visual field. Here, we follow previous evidence acquired at 7T that suggested, for the first time, that perceptual long-range integration of bistable motion is linked to the interhemispheric functional connectivity in hMT+ (Sousa et al., \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). However, that finding remains the single one in the literature suggesting that interhemispheric perceptual integration is mirrored by enhanced functional connectivity across the hemispheres Specifically, the working hypothesis is that interhemispheric functional connectivity hMT+, at the scale of fMRI temporal resolution, reflects perceptual decision (eg. Coherence vs Incoherence). Using a different bistable stimulus (moving plaids, Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e) and a 3T MRI scanner, we aimed to validate these findings at different fMRI temporal resolutions, which depend on the acquisition Repetition Time (TR). Our motivation lies in the ability to detect fast perceptual transitions, in the order of magnitude of the TR, to then explore its neuroscientific potential as well as the possibility of using this information in a neurofeedback setup. We use a moving plaid, which is a bistable visual stimulus that can be seen as two gratings sliding over each other and moving in different directions (incoherent motion) or as a single surface moving coherently – the two gratings are integrated perceptually into a single surface moving coherently in a direction intermediate to the directions of the individual gratings (coherent motion) (Sousa et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2018\u003c/span\u003e, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Sayal et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e)\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eMultiband echo-planar imaging (MB-EPI) sequences have revolutionized image sampling rates, offering a remarkable increase in temporal resolution. These use advanced pulse sequences and specialized MRI hardware to simultaneously excite and acquire data from multiple slices. By doing so, they can cover a larger volume of the brain or body in a shorter time compared to the sequential acquisition of slices. In clinical applications, these sequences show more detailed BOLD signals with superior temporal resolution and contrast-to-noise ratio, pivotal features for detecting nuanced responses of neuronal populations (Zhang et al., \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). However, high multiband factors should be used with caution, as discussed in a recent editorial (Wall, \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eGiven the increasing application of MB sequences for fMRI protocols, a number of studies started to thoroughly evaluate how the BOLD signal was affected by this new acquisition technique (Chen et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Todd et al., \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2016\u003c/span\u003e, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Demetriou et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Risk, Kociuba and Rowe, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Jahanian et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Bhandari et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Renz et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). These studies compare quality and functional metrics for data acquired with different temporal resolutions (hence multiband factors). For instance, (Chen et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2015\u003c/span\u003e) results revealed notably higher BOLD information content when employing faster TRs between 300 ms and 600 ms, in contrast to a 2-second TR, indicating that faster TRs can capture more information per unit of time in task-based fMRI studies. Another study (Todd et al. \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2016\u003c/span\u003e) evaluated BOLD sensitivity and false-positive activations. Based on the results, the authors make recommendations regarding the reconstruction method and the MB factor that allows for a combined low probability of false-positive activations (due to slice leakage) and high quality (fair spatial and temporal resolution) data. Their conservative recommendation limits the MB factor to 2 and in-plane acceleration to 2. In fact, the benefits of these sequences are not always clear. (Demetriou et al. \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) found that multi-band protocols showed strong benefits for resting-state data analyses and Multi-Voxel Pattern Analyses (MVPA) but not so evident and consistent improvements for task-based paradigms and standard General Linear Model (GLM) approaches, linking these results to variations in temporal signal-to-noise-ratio (SNR) between sequences. (Srirangarajan et al., \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) provided a cautionary note regarding the study of mesolimbic regions with multiband sequences, as the authors found reductions in temporal SNR that could account for impaired detection of task-related responses in these deeper regions in the brain.\u003c/p\u003e \u003cp\u003eNeurofeedback (NF) experiments based on real-time fMRI have opened the possibility for individuals to self-regulate specific not only brain activation but now also connectivity patterns. This is achieved by interpreting a feedback signal and exploring the appropriate strategies for its modulation. Currently, the impact of several protocol parameters with special attention given to exploring the influence of different self-regulation strategies, feedback interfaces, and data analysis pipelines (Sitaram et al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Hampson, Ruiz and Ushiba, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Ros et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Weber, Ethofer and Ehlis, \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). As was recently demonstrated (Kadosh and Staunton, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), psychological variables such as motivation or mood correlate with the success of neurofeedback.\u003c/p\u003e \u003cp\u003eWhen feedback is presented in a continuous manner, it requires updating the feedback signal after each fMRI time point. Considering a TR for functional imaging of 2 seconds, we achieve a temporal resolution of 0.5 Hz in the feedback signal. This rather slow update rate of the feedback interface, coupled with the hemodynamic delay between the self-regulation and the feedback, may reduce the efficacy of the protocols. While decreasing the TR does not influence the rate at which neural processes occur nor the response time of the HRF, it may positively impact the immersiveness and the feedback calculation estimates for real-time use. Specifically, this increased temporal resolution could play a role in improving neurofeedback success, as it may contribute to the enhanced reliability of feedback signals based on activation or connectivity measures, providing a more realistic and immersive experience.\u003c/p\u003e \u003cp\u003eIn this study, we investigated our working hypothesis of interhemispheric functional connectivity in hMT + during a bistable visual motion task at four temporal resolutions: 2, 1.33, 1, and 0.4 Hz (TR = 0.5, 0.75, 1, and 2.5 seconds, respectively). This led to confirmation that one can link interhemispheric functional connectivity to global coherent perception. Then, we test if the activation metrics of hMT + are replicable across TRs. Finally, by establishing that perception can be read out from connectiviy patterns, we compare and contrast the characteristics of the feedback signal to be provided in a neurofeedback loop based on the activation and connectivity of hMT+. This region has been used in past studies as the target for neurofeedback, aiming for the self-modulation of the activity of this region in real time (Banca et al., \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Sousa et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Direito et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). If we aimed to provide feedback on the perceptual switches based on the hMT + inter-hemispheric correlation, we checked how reliable and responsive this feedback would be at different resolutions.\u003c/p\u003e "},{"header":"Methods","content":"\u003cp\u003eParticipants\u003c/p\u003e\u003cp\u003eFifteen healthy participants were recruited for this experiment (mean age 29.7 ± 8.4 years, 6 females) with normal or corrected-to-normal vision and no history of neurological or psychiatric diseases. All participants were right-handed, as confirmed by a handedness questionnaire adapted from (Oldfield, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e1971\u003c/span\u003e): mean laterality index of 85.0 ± 9.0. All gave informed written consent before participating in accordance with the declaration of Helsinki, and the study followed the safety guidelines for magnetic resonance imaging research on humans. The work was approved by the Ethics Committee of the Faculty of Medicine of the University of Coimbra.\u003c/p\u003e\u003cp\u003efMRI data acquisition\u003c/p\u003e\u003cp\u003eScanning was performed on a 3T Siemens Magnetom Prisma fit, using a 64-channel head/neck coil, at the Institute of Nuclear Sciences Applied to Health (ICNAS), University of Coimbra, Portugal. The scanning session started with the acquisition of one 3D anatomical magnetization-prepared rapid acquisition gradient echo (MPRAGE) pulse sequence (TR = 2530 ms, echo time (TE) = 3.42 ms, flip angle (FA) = 7°, 176 slices, voxel size 1.0 × 1.0 × 1.0 mm\u003csup\u003e3\u003c/sup\u003e, field of view (FOV) = 256 × 256 mm\u003csup\u003e2\u003c/sup\u003e).\u003c/p\u003e\u003cp\u003eThe functional runs were acquired using a 2D multi-band (MB) gradient-echo (GE) echo-planar imaging (EPI) sequence from the Center for Magnetic Resonance Research, University of Minnesota (Release R016a). We tested four different temporal resolutions: TR = 0.5 s (MB factor = 6, FA = 53º, 42 slices), TR = 0.75 s (MB factor = 4, FA = 63º, 40 slices), TR = 1 s (MB factor = 3, FA = 68º, 42 slices), TR = 2.5 s (MB factor = 1, FA = 85º, 42 slices). The remaining parameters were matched: TE = 30.2 ms, interleaved slices with 0.5 mm gap, voxel size 2.5 × 2.5 × 2.5 mm\u003csup\u003e3\u003c/sup\u003e, FoV 192 × 192 mm\u003csup\u003e2\u003c/sup\u003e (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e \u003c/p\u003e\u003cdiv class=\"gridtable\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eMain parameters of the four fMRI sequences.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e\u003ccolgroup cols=\"5\"\u003e\u003c/colgroup\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSequence\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eTR (s)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.5\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.75\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2.5\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eTE (ms)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colspan=\"4\" nameend=\"c5\" namest=\"c2\"\u003e \u003cp\u003e30.2\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eMB factor\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eVoxel size (mm)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colspan=\"4\" nameend=\"c5\" namest=\"c2\"\u003e \u003cp\u003e2.5\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSlice gap (mm)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colspan=\"4\" nameend=\"c5\" namest=\"c2\"\u003e \u003cp\u003e0.5\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eNumber of slices\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e42\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e40\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e42\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e42\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eFlip Angle (º)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e53\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e63\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e68\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e85\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eBandwidth (Hz/px)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2742\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2742\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2632\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1994\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eEcho Spacing (ms)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.51\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.49\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.49\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.59\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eExcite pulse duration (us)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2560\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2560\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2560\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4840\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eEPI factor\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colspan=\"4\" nameend=\"c5\" namest=\"c2\"\u003e \u003cp\u003e76\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eGRAPPA\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOff\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eOff\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eOff\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eLeakBlock\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOn\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eOn\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eOn\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eOff\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/table\u003e\u003c/div\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eFor mapping and correction of image distortions related to magnetic field inhomogeneities, we acquired a pair of spin-echo images with anterior-posterior (AP) and posterior-anterior (PA) phase encoding polarity with matching geometry and echo-spacing to each of the functional scans. These were acquired before the functional runs.\u003c/p\u003e\u003cp\u003eThe participants’ physiological signals (respiration and pulse) were recorded during the functional runs using the scanner’s Physiological Measurement Unit (PMU). The respiratory signal was recorded at 50 Hz using a respiratory cushion, and the cardiac cycle was recorded at 200 Hz using a pulse sensor.\u003c/p\u003e\u003cp\u003eFunctional tasks\u003c/p\u003e\u003cp\u003eWe implemented three functional tasks based on a moving plaid stimulus. These stimuli are created by superimposing two gratings leading to a bistable percept: stimuli can be perceived moving coherently as a single surface (integration) or incoherently as two separate surfaces sliding over each other (segregation) (see supplementary materials and Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe first is a localizer for the region of interest - hMT+. For this purpose, we created three conditions: i. ‘Fixation’ - a fixation cross; ii. ‘Static plaid’ - a stationary plaid; iii. ‘moving plaid’ - a moving plaid, which is inherently a bistable stimulus (coherent versus incoherent, e.g. integration versus segregation). The run lasted for 2.9 min and was composed of nine trials with the sequential presentation of each condition for 6 seconds.\u003c/p\u003e\u003cp\u003eWe named the following two tasks as ‘ambiguous’ and ‘unambiguous’ runs. These runs are composed of trials considering three conditions: ‘static’ (static plaid), ‘motion’ (ambiguous or unambiguous moving plaid), and ‘MAE’ (a period during which motion aftereffect is expected). During the ambiguous runs, the ‘motion’ condition showed the participants the moving plaid without any overlaid dots. As such, the stimulus is entirely ambiguous (the percept alternates between coherent and incoherent). Here, the participants were instructed to report the perceived type of motion (coherent or incoherent) using two buttons of a response box. During the unambiguous runs, the plaid is shown with overlaid dots moving either coherently down or incoherently inwards, which disambiguates the perception of the plaid (unambiguously coherent or incoherent). Based on the responses given by the participant in the previous ambiguous run, we manipulated the switches between coherent and incoherent motion in the unambiguous runs to match the previous responses precisely. This matched the time of each percept across both types of runs. The participants received the same instruction - to report the perceived type of motion at all times.\u003c/p\u003e\u003cp\u003eIn this work, we considered the localizer and the four unambiguous runs, one for each temporal resolution.\u003c/p\u003e\u003cp\u003efMRI data preprocessing\u003c/p\u003e\u003cp\u003eThe data were organized according to the Brain Imaging Data Structure (BIDS) (Gorgolewski et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2016\u003c/span\u003e), using \u003cem\u003eBIDSkit\u003c/em\u003e (Tyszka, \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) and \u003cem\u003edcm2niix\u003c/em\u003e (Li et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). Results included in this manuscript were obtained after data preprocessing performed using \u003cem\u003efMRIPrep\u003c/em\u003e 23.0.2 (Esteban et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), which is based on \u003cem\u003eNipype\u003c/em\u003e 1.8.6 (Gorgolewski et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2011\u003c/span\u003e; Esteban et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). For a detailed description of the fmriPrep pipeline please see supplementary materials. To perform quality checks on the anatomical and functional images and extract metrics for further analysis and sequence comparison we ran \u003cem\u003emriqc\u003c/em\u003e 23.0.1 (Esteban et al., \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). All subsequent analyses were performed in Python using Nilearn (Abraham et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2014\u003c/span\u003e).\u003c/p\u003e\u003cp\u003efMRI data analysis\u003c/p\u003e\u003cp\u003eRegarding the localizer data, hMT + was functionally localized for each subject using a standard GLM analysis. The design matrix included predictors for all experimental conditions (‘fixation’, ‘static plaid', ‘moving plaid’) and confound regressors based on the mean and first derivative of the voxels in the CSF and the six head motion parameters and their derivatives. Temporal high-pass filtering (cut-off = 0.03 Hz) and spatial smoothing (FWHM = 6 mm) were applied. hMT + was manually selected on both hemispheres based on the activation map for the contrast ‘moving plaid \u0026gt; static plaid’ corrected with Bonferroni p = 0.05, cluster threshold k = 50, and confirmed using the hMT + mask of Neurosynth (Yarkoni et al., \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2011\u003c/span\u003e)). Then, a spherical mask was designed around the center coordinates of each defined ROI (radius = 6 mm).\u003c/p\u003e\u003cp\u003eThe first step to analyzing the unambiguous runs was the definition of the coherent and incoherent predictors. To this end, using the perceptual reports acquired via button presses with a temporal resolution of 60 Hz, we calculated the percentage of coherent reports over the total number of reports for each volume. We used this value to label each volume as coherent or incoherent. This information defines which of the two percepts the participant is experiencing at any given time, allowing us to design predictors for coherent and incoherent events (for the GLM analyses) and to have the onsets and offsets of each percept to use for the neurofeedback algorithm.\u003c/p\u003e\u003cp\u003eWe estimated the functional activation maps of the unambiguous runs for each TR using a GLM. The design matrix included predictors for all experimental conditions (‘static’, ‘unambiguous motion’, ‘MAE’) and confound regressors based on the mean and first derivative of the voxels in the CSF and the six head motion parameters and their derivatives. Temporal high-pass filtering (cut-off = 0.003 Hz) was applied before extracting ROI activation measures (beta statistic and t-value). Our contrast of interest here was ‘unambiguous motion’ vs. ‘static’.\u003c/p\u003e\u003cp\u003eWe then used a different design matrix, replacing the ‘unambiguous motion’ predictor with predictors for coherent and incoherent percepts. As in the previous model, we extract the ROI beta statistic and t-value for the contrasts between coherent/incoherent and static plaid.\u003c/p\u003e\u003cp\u003eFor the following analyses, we extract the time course of the ROI and normalize it as a percent signal change to the mean value across time. We estimate the feedback based on this time course as we would in an actual neurofeedback experiment, by displaying the signal variation during the ‘unambiguous motion’ upregulation condition vs. the ‘static’ baseline/downregulation condition.\u003c/p\u003e\u003cp\u003eNext, we use the time courses of the left and right hMT + to estimate Pearson’s correlation over time, a measure of inter-hemispheric functional connectivity. We use a sliding window of 6 seconds for the three lower TRs and of 7.5 seconds for the higher TR (a minimum of 3 data points was considered for obtaining a measurable correlation value). Using the perceptual report information, we center all correlation windows at the transitions between coherent and incoherent percepts and average them across trials and participants. Given the size of the windows, we only considered coherent and incoherent events that lasted at least 7.5 seconds.\u003c/p\u003e\u003cp\u003eTo use the functional connectivity information as feedback, we study its characteristics by defining a feedback rule. We defined that successful automatic detection of a perceptual transition meant a 10% minimum increase or decrease in correlation, for coherent and incoherent trials, respectively. We chose this threshold based on (Sousa et al., \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), which reported an average difference in correlation between percepts of 0.13 ± 0.04. Based on this rule, we extract the transitions’ detection ratio and the time to detection from the first correlation window that includes the volume of transition.\u003c/p\u003e\n\u003ch2\u003eData and code availability\u003c/h2\u003e\n\u003cp\u003eAll the code for the above-mentioned analyses and data (including the activation maps and hMT+ timecourses) can be found at https://github.com/alexsayal/vpmb-tr.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eLocalizer for hMT+\u003c/p\u003e \u003cp\u003eWe defined the left and right hMT\u0026thinsp;+\u0026thinsp;ROIs in the localizer map of each individual, considering the contrast \u0026lsquo;moving\u0026rsquo; vs. \u0026lsquo;static\u0026rsquo;, a voxel-wise correction with Bonferroni\u0026rsquo;s method (p\u0026thinsp;=\u0026thinsp;0.05) and cluster-wise with a minimum cluster size of 50 voxels (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). The MNI coordinates of the center of the left and right hMT\u0026thinsp;+\u0026thinsp;clusters for each subject are reported in Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e. The group activation map is shown in Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e, and the average MNI coordinates of the center of the left and right clusters are (-42,-72,4) and (48,-66,6), respectively.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eGLM statistical values of hMT+\u003c/p\u003e \u003cp\u003eIn Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, we display the outcomes of the GLM analyses performed. In the first column, we plot the beta and t-values of the contrast \u0026lsquo;motion vs. static\u0026rsquo;. While there is no difference in the beta values between TRs, the t-value of the sequence with the lowest TR is significantly higher than the 2.5 s sequence. This indicates a higher level of certainty in the activation measurement (lower beta error) for the sequence with TR\u0026thinsp;=\u0026thinsp;0.5 s.\u003c/p\u003e \u003cp\u003eWith the coherent and incoherent contrasts, we aimed to test if the measures of activation of rapidly changing events were higher and more reliable for the higher-resolution data. We found a significant difference in the t-value of the coherent contrast, again between the sequences with TR of 0.5 and 2.5 seconds.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eInterhemispheric connectivity\u003c/p\u003e \u003cp\u003eIn Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e, we display the time courses of interhemispheric correlation of hMT\u0026thinsp;+\u0026thinsp;for all the sequences with different temporal resolutions, centered on the perceptual switches between coherent and incoherent states and vice versa. The results are consistent across TRs - the transition to the coherent percept is linked to an increase in interhemispheric connectivity, while the transition to the incoherent percept is linked to a decrease in connectivity.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eBOLD feedback estimation based on the hMT\u0026thinsp;+\u0026thinsp;response\u003c/p\u003e \u003cp\u003eWe display the group average hMT\u0026thinsp;+\u0026thinsp;time course for each TR in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e. This allows the visual comparison between the signals - the difference in temporal resolution is clear while the overall signal variation is similar across TRs. Using these time courses, we simulate a feedback signal to be provided during the motion/upregulation condition. We first normalize the signal to the mean value of the \u0026lsquo;static\u0026rsquo;, baseline condition, as a percent signal variation. Then, based on a maximum percent signal change of 2.5, we discretize this signal into 10 values, as we would if this information were to be displayed in a visual feedback interface with 10 activation levels. We display both of these signals in Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e for the four TRs.\u003c/p\u003e \u003cp\u003eWhile all temporal resolutions provide a valid feedback signal, the variability for the 2.5 s sequence is higher (average SD\u0026thinsp;=\u0026thinsp;4.97, while for the other sequences, it ranges from 1.39 to 1.60), with the feedback signal achieving values of zero during the upregulation condition. All signals take approximately the same time to reach the peak value of activation (as indicated by the red vertical dashed line).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFeedback characteristics based on interhemispheric functional connectivity\u003c/p\u003e \u003cp\u003eHere, we explored two characteristics of the feedback signal based on interhemispheric hMT\u0026thinsp;+\u0026thinsp;functional connectivity at different resolutions. Based on a rule that could distinguish coherent or incoherent perception, we calculated the percentage of detected switches (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e) and the time it took to detect such switches (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e) since the volume of transition for each sequence. The results show statistically significant differences in both metrics between the sequence with TR of 2.5 seconds and the other three sequences.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eHere we found that changes in interhemispheric connectivity reliably reflects perceptual integration and decision and becomes more evident with improved temporal resolution as implemented by multiband sequences. We also showed the feasibility of using connectivity based neurofeedback approaches\u003c/p\u003e \u003cp\u003eImproving temporal resolution as a prior to study connectivity\u003c/p\u003e \u003cp\u003eHere we used moving plaid stimuli, that lead to bistable percepts of coherent (reflecting integration) or non coherent (reflecting segregation) motion which lead to different activity levels in visual area hMT+, which may relate to differential profiles of adaptation and inhibition strengths across time for the two percepts (Sousa et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2018\u003c/span\u003e, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Sayal et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eOur first aim was to compare activation metrics across temporal resolutions for each of the conditions, coherent and incoherent motion percepts. The hypothesis was that higher temporal resolution sequences would allow for better activation estimates (i.e. smaller beta error values), which may be critical for subsequent connectivity analysis. Indeed, we found a significant difference between the t-values of the lower and higher temporal resolution sequences for the overall motion and coherent predictors, with higher t-values for the lower TR. This result indicates that higher resolution sequences might benefit rapid-changing events or short blocks, as they provide more points to estimate the GLM and improve the statistical outcomes. These results might, however, be specific to the research question, as (Dar\u0026aacute;nyi et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) found that increasing the temporal resolution of the signal did not universally improve the group-level statistical outcomes, reinforcing the importance of pilot testing and hemodynamic response function estimation for the ROIs being studied.\u003c/p\u003e \u003cp\u003eFunctional interhemispheric connectivity reliably reflects perceptual coherence\u003c/p\u003e \u003cp\u003eA recent study emphasized the crucial role of replication by demonstrating that a meta-analytical approach, pooling information across teams, was essential in establishing a significant consensus on activated regions (Rotem Botvinik-Nezer \u003cem\u003eet al.\u003c/em\u003e, 2020). We previously suggested using roof like stimuli at 7 tesla that interhemispheric functional connectivity reflects perceptual integration (Sousa et al., \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Remarkably this hypothesis lacks generalization. Here we confirmed this hypothesis using a different bistable stimulus at a lower field strength and distinct temporal resolutions.\u003c/p\u003e \u003cp\u003eThe results of (Sousa et al., \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) showed opposing variations in hMT\u0026thinsp;+\u0026thinsp;inter-hemispheric correlation when changing either to coherent or incoherent percepts of an ambiguous visual stimulus. The connectivity between the left and right hMT\u0026thinsp;+\u0026thinsp;was shown to be critical in interpreting visual moving stimuli. When studying the \u0026ldquo;motion quartet\u0026rdquo;, an ambiguous stimulus that generates horizontal or vertical apparent motion, it was suggested that its perception requires interhemispheric integration, although functional connectivity was not studied (Gen\u0026ccedil; et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). The authors were able to predict the perception based on structural connectivity data of callosal segments linking both hMT\u0026thinsp;+\u0026thinsp;regions. Howver these authors did not present functional connectivity data. Nevertheles, when the connectivity is manipulated using transcranial magnetic stimulation, the sensitivity to one of the possible percepts increases (Chiappini et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e When we analyze the volumes around the transition between coherent and incoherent perceptual states, as reported by the participants, we found a decrease in inter-hemispheric correlation after switching to the incoherent percept and an increase after switching to the coherent percept (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e), which is consistent with the findings of (Sousa et al., \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) (see Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e of that study). One should note that this pioneering study uses a different ambiguous stimulus and a scanner with a higher static magnetic field (7T), but reaches consistent conclusions regarding this visual region activation when processing ambiguous moving stimuli. Moreover, in this study, the findings are replicated for all TRs.\u003c/p\u003e \u003cp\u003eFunctional connectivity as feedback of perceptual transitions\u003c/p\u003e \u003cp\u003eWe estimate the impact of temporal resolution on a BOLD-based feedback signal in Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e. An intriguing finding is that, for the sequence with the lowest resolution, the variability of the feedback signal is much higher. The standard deviation indicates that the feedback signal is less consistent across participants and that the mean even reaches zero at some points along the \u0026lsquo;upregulation\u0026rsquo; block. This was not the case for the sequences with TR of 0.5 and 0.75 seconds. We found no difference in the time it takes for the signal from each sequence to reach the peak value of activation.\u003c/p\u003e \u003cp\u003eHaving established that the modulation of a single brain region in neurofeedback studies has an impact not only in the target region but also on a network of regions, neuroimaging studies move towards studying the connectivity between nodes of these networks (Ramot and Martin, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). In this sense, if one aims to use neurofeedback to target a specific function or behavior, one should characterize the functional network involved and the appropriate target region, assess if the changes we may be inducing are propagating along the rest of the network, or even if we should be targeting some parameter of the network itself.\u003c/p\u003e \u003cp\u003eHere, we explore the interhemispheric functional connectivity of a region that responds to our stimuli, assessing its potential as a target for neuromodulation at different TRs. This has been successfully applied before in other neurofeedback paradigms (Pereira et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Wang et al., \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Hayashi et al., \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Specifically, we simulated the feedback of the perceptual switches based on the hMT\u0026thinsp;+\u0026thinsp;interhemispheric correlation value. We found that the feedback information is both more reliable and more responsive for the higher resolution sequences (Figs.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e and \u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e) when we consider a simple rule for determining a perceptual switch. Based on a previous study, which found that the more information available to the participant during neurofeedback, the greater its ability to self-regulate brain activity (Sorger et al., \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2018\u003c/span\u003e), we conclude that the time delay between the modulation and the feedback of that modulation time is an important factor in neurofeedback paradigms. As such, if the goal is to provide feedback based on connectivity, one could consider using higher temporal resolution sequences.\u003c/p\u003e \u003cp\u003eTrade-offs, limitations and compromises\u003c/p\u003e \u003cp\u003eThe neural processes indirectly inferred using fMRI vary in the scale of the millisecond but are sampled in the scale of the second. This represents a severe downsampling with fMRI sample intervals typically ranging from 1 to 3 s, way longer than typical interneuron delays (Seth, Chorley and Barnett, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). Nevertheless, this measure of neuronal activity can be interpreted as a representation of a low pass filter of the local field potentials, showing that neural patterns can also be found in the BOLD signal, a particularly relevant feature for connectivity methods applied to fMRI data (Fernandes et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). While the temporal resolution of this neuroimaging technique is far from others, such as electroencephalography, lowering the TR can provide us with a better sampling of the HRF and more certainty when studying brain connectivity with functional MRI.\u003c/p\u003e \u003cp\u003eIncreasing the temporal resolution during real-time acquisitions poses challenges both in image quality and technical implementations. As recent research suggests, multiband sequences may diminish the detection ability of activity in mesolimbic regions (Srirangarajan et al., \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) but not in other regions in the brain, as we have seen here with hMT+. Also, the temporal SNR is not homogeneous in the whole brain and tends to be lower with higher multiband factors (Todd et al., \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Srirangarajan et al., \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Hence, the recommendation for the use of these sequences for real-time fMRI is highly dependent on the target region or network of interest.\u003c/p\u003e \u003cp\u003eFrom the technical point of view, fMRI neurofeedback has high computational requirements, namely regarding image reconstruction at the scanner, network capacity and speed for image transfer at every volume, and the real-time image processing software itself. Some of these requirements have been addressed in (Renz et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), where the authors managed to achieve a stable real-time acquisition with a TR of 1 second, but not less, a result which may be highly dependent on each site\u0026rsquo;s MRI setup and infrastructure.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eIn conclusion, our study generalizes the hypothesis that interhemispheric connectivity reflects spontaneous perceptual switches in perceptual coherence, which is shown across temporal resolutions (even at 3T). We found stable activation metrics in the hMT\u0026thinsp;+\u0026thinsp;region across resolutions and consistent results of changes in interhemispheric connectivity as a function of perceptual decision, affirming the reliability of fMRI for studying dynamic connectivity patterns. While simulating a feedback signal based on interhemispheric connectivity, higher resolution sequences (TRs within 1 second and lower) showed benefits. Overall, integrating fast fMRI sequences may offer promising avenues for optimizing cognitive functions within neurofeedback settings.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eData and code availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll the code for the above-mentioned analyses and data (including the activation maps and hMT+ timecourses) can be found at https://github.com/alexsayal/vpmb-tr.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions \u003c/strong\u003e\u0026ndash; AS, TS and MCB designed the study ,BD, TS and MCB provided supervision, AS acquired and analysed the data, SA acquired the data, AS wrote the first draft of the manuscript, all authors revised, edited and approved the final version.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u0026nbsp;\u0026ndash; the authors report no competing interests\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e: This research work was funded by the Portuguese Foundation for Science and Technology (FCT) (grants: UID/04950B/2020, UID/04950P/2020, DSAIPA/DS/0041/2020, PTDC/PSI-GER/1326/2020, \u0026nbsp;2022.02963.PTDC \u0026ndash; Hallucin and by the BIAL Foundation project 207/16.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAbraham, A., Pedregosa, F., Eickenberg, M., Gervais, P., Mueller, A., Kossaifi, J., Gramfort, A., Thirion, B. and Varoquaux, G. (2014) \u0026lsquo;Machine learning for neuroimaging with scikit-learn\u0026rsquo;, \u003cem\u003eFrontiers in Neuroinformatics\u003c/em\u003e, 8. Available at: https://doi.org/10.3389/fninf.2014.00014.\u003c/li\u003e\n\u003cli\u003eBanca, P., Sousa, T., Catarina Duarte, I. and Castelo-Branco, M. (2015) \u0026lsquo;Visual motion imagery neurofeedback based on the hMT+/V5 complex: evidence for a feedback-specific neural circuit involving neocortical and cerebellar regions\u0026rsquo;, \u003cem\u003eJournal of Neural Engineering\u003c/em\u003e, 12(6), p. 066003. Available at: https://doi.org/10.1088/1741-2560/12/6/066003.\u003c/li\u003e\n\u003cli\u003eBhandari, R., Kirilina, E., Caan, M., Suttrup, J., De Sanctis, T., De Angelis, L., Keysers, C. and Gazzola, V. (2020) \u0026lsquo;Does higher sampling rate (multiband + SENSE) improve group statistics - An example from social neuroscience block design at 3T\u0026rsquo;, \u003cem\u003eNeuroImage\u003c/em\u003e, 213, p. 116731. Available at: https://doi.org/10.1016/j.neuroimage.2020.116731.\u003c/li\u003e\n\u003cli\u003eChen, L., T. Vu, A., Xu, J., Moeller, S., Ugurbil, K., Yacoub, E. and Feinberg, D.A. (2015) \u0026lsquo;Evaluation of highly accelerated simultaneous multi-slice EPI for fMRI\u0026rsquo;, \u003cem\u003eNeuroImage\u003c/em\u003e, 104, pp. 452\u0026ndash;459. Available at: https://doi.org/10.1016/j.neuroimage.2014.10.027.\u003c/li\u003e\n\u003cli\u003eChiappini, E., Sel, A., Hibbard, P.B., Avenanti, A. and Romei, V. (2022) \u0026lsquo;Increasing interhemispheric connectivity between human visual motion areas uncovers asymmetric sensitivity to horizontal motion\u0026rsquo;, \u003cem\u003eCurrent Biology\u003c/em\u003e, 32(18), pp. 4064-4070.e3. Available at: https://doi.org/10.1016/j.cub.2022.07.050.\u003c/li\u003e\n\u003cli\u003eDar\u0026aacute;nyi, V., Hermann, P., Homolya, I., Vidny\u0026aacute;nszky, Z. and Nagy, Z. (2021) \u0026lsquo;An empirical investigation of the benefit of increasing the temporal resolution of task-evoked fMRI data with multi-band imaging\u0026rsquo;, \u003cem\u003eMagnetic Resonance Materials in Physics, Biology and Medicine\u003c/em\u003e, 34(5), pp. 667\u0026ndash;676. Available at: https://doi.org/10.1007/s10334-021-00918-z.\u003c/li\u003e\n\u003cli\u003eDemetriou, L., Kowalczyk, O.S., Tyson, G., Bello, T., Newbould, R.D. and Wall, M.B. (2018) \u0026lsquo;A comprehensive evaluation of increasing temporal resolution with multiband-accelerated protocols and effects on statistical outcome measures in fMRI\u0026rsquo;, \u003cem\u003eNeuroImage\u003c/em\u003e, 176, pp. 404\u0026ndash;416. Available at: https://doi.org/10.1016/j.neuroimage.2018.05.011.\u003c/li\u003e\n\u003cli\u003eDireito, B., Ramos, M., Pereira, J., Sayal, A., Sousa, T. and Castelo-Branco, M. (2021) \u0026lsquo;Directly Exploring the Neural Correlates of Feedback-Related Reward Saliency and Valence During Real-Time fMRI-Based Neurofeedback\u0026rsquo;, \u003cem\u003eFrontiers in Human Neuroscience\u003c/em\u003e, 14, p. 578119. Available at: https://doi.org/10.3389/fnhum.2020.578119.\u003c/li\u003e\n\u003cli\u003eDuarte, J.V., Costa, G.N., Martins, R. and Castelo‐Branco, M. (2017) \u0026lsquo;Pivotal role of hMT+ in long‐range disambiguation of interhemispheric bistable surface motion\u0026rsquo;, \u003cem\u003eHuman Brain Mapping\u003c/em\u003e, 38(10), pp. 4882\u0026ndash;4897. Available at: https://doi.org/10.1002/hbm.23701.\u003c/li\u003e\n\u003cli\u003eEsteban, O., Birman, D., Schaer, M., Koyejo, O.O., Poldrack, R.A. and Gorgolewski, K.J. (2017) \u0026lsquo;MRIQC: Advancing the automatic prediction of image quality in MRI from unseen sites\u0026rsquo;, \u003cem\u003ePLOS ONE\u003c/em\u003e. Edited by B.C. Bernhardt, 12(9), p. e0184661. Available at: https://doi.org/10.1371/journal.pone.0184661.\u003c/li\u003e\n\u003cli\u003eEsteban, O., Markiewicz, C.J., Burns, C., Goncalves, M., Jarecka, D., Ziegler, E., Berleant, S., Ellis, D.G., Pinsard, B., Madison, C., Waskom, M., Notter, M.P., Clark, Daniel, Manh\u0026atilde;es-Savio, A., Clark, Dav, Jordan, K., Dayan, M., Halchenko, Y.O., Loney, F., Norgaard, M., Salo, T., Dewey, B.E., Johnson, H., Bougacha, S., Keshavan, A., Yvernault, B., Hamalainen, C., Christian, H., Ćirić, R., Dubois, M., Joseph, M., Cipollini, B., Tilley II, S., Visconti di Oleggio Castello, M., De La Vega, A., Wong, J., Kaczmarzyk, J., Huntenburg, J.M., Clark, M.G., Kent, J.D., Benderoff, E., Erickson, D., Dias, M. de F., Hanke, M., Giavasis, S., Moloney, B., Nichols, B.N., Tungaraza, R., Dell\u0026rsquo;Orco, A., Frohlich, C., Wassermann, D., de Hollander, G., Koudoro, S., Eshaghi, A., Millman, J., Mancini, M., Close, T., Nielson, D.M., Varoquaux, G., Waller, L., Watanabe, A., Mordom, D., Guillon, J., Robert-Fitzgerald, T., Chetverikov, A., Rokem, A., Acland, B., Forbes, J., Markello, R., Gillman, A., Bernardoni, F., Kong, X.-Z., Geisler, D., Salvatore, J., Gramfort, A., Doll, A., Buchanan, C., DuPre, E., Liu, S., Schaefer, A., Kleesiek, J., Sikka, S., Schwartz, Y., Ghayoor, A., Lee, J.A., Mattfeld, A., Richie-Halford, A., Liem, F., Vaillant, G., Perez-Guevara, M.F., Heinsfeld, A.S., Haselgrove, C., Durnez, J., Lampe, L., Poldrack, R., Glatard, T., Baratz, Z., Tabas, A., Cumba, C., P\u0026eacute;rez-Garc\u0026iacute;a, F., Blair, R., Iqbal, S., Welch, D., Ben-Zvi, G., Contier, O., Triplett, W., Craddock, R.C., Correa, C., Papadopoulos Orfanos, D., Stadler, J., Warner, J., Sisk, L.M., Falkiewicz, M., Sharp, P., Rothmei, S., Kim, S., Weinstein, A., Kahn, A.E., Kastman, E., Bottenhorn, K., Grignard, M., Perkins, L.N., Zhou, D., Bielievtsov, D., Cooper, G., Stojic, H., Hui Qian, T., Linkersd\u0026ouml;rfer, J., Renfro, M., Hinds, O., Stanley, O., K\u0026uuml;ttner, R., Pauli, W.M., Xie, X., Glen, D., Kimbler, A., Meyers, B., Tarbert, C., Ginsburg, D., Haehn, D., Margulies, D.S., Condamine, E., Ma, F., Malone, I.B., Snoek, L., Brett, M., Cieslak, M., Hallquist, M., Molina-Romero, M., Bilgel, M., Lee, N., Kuntke, P., Jalan, R., Inati, S., Gerhard, S., Mathotaarachchi, S., Saase, V., Van, A., Steele, C.J., Ort, E., Lerma-Usabiaga, G., Schwabacher, I., Arias, J., Lai, J., Pellman, J., Huguet, J., WEN, J., Leinweber, K., Chawla, K., Weninger, L., Modat, M., Mukhometzianov, R., Harms, R., Andberg, S.K., Matsubara, K., Gonz\u0026aacute;lez Orozco, A.A., Routier, A., Marina, A., Davison, A., Floren, A., Park, A., Frederick, B., Cheung, B., McDermottroe, C., McNamee, D., Shachnev, D., Vogel, D., Flandin, G., Jones, H., Gonzalez, I., Varada, J., Schlamp, K., Podranski, K., Huang, L., Noel, M., Pannetier, N., Numssen, O., Khanuja, R., Urchs, S., Shim, S., Nickson, T., Huang, L., Broderick, W., Tambini, A., Mihai, P.G., Gorgolewski, K.J. and Ghosh, S. (2022) \u0026lsquo;nipy/nipype: 1.8.3\u0026rsquo;. Zenodo. Available at: https://doi.org/10.5281/ZENODO.596855.\u003c/li\u003e\n\u003cli\u003eEsteban, O., Markiewicz, C.J., Goncalves, M., Provins, C., Kent, J.D., DuPre, E., Salo, T., Ciric, R., Pinsard, B., Blair, R.W., Poldrack, R.A. and Gorgolewski, K.J. (2023) \u0026lsquo;fMRIPrep: a robust preprocessing pipeline for functional MRI\u0026rsquo;. Zenodo. Available at: https://doi.org/10.5281/ZENODO.852659.\u003c/li\u003e\n\u003cli\u003eFernandes, T.T., Direito, B., Sayal, A., Pereira, J., Andrade, A. and Castelo-Branco, M. (2020) \u0026lsquo;The boundaries of state-space Granger causality analysis applied to BOLD simulated data: A comparative modelling and simulation approach\u0026rsquo;, \u003cem\u003eJournal of Neuroscience Methods\u003c/em\u003e, 341, p. 108758. Available at: https://doi.org/10.1016/j.jneumeth.2020.108758.\u003c/li\u003e\n\u003cli\u003eGen\u0026ccedil;, E., Bergmann, J., Singer, W. and Kohler, A. (2011) \u0026lsquo;Interhemispheric Connections Shape Subjective Experience of Bistable Motion\u0026rsquo;, \u003cem\u003eCurrent Biology\u003c/em\u003e, 21(17), pp. 1494\u0026ndash;1499. Available at: https://doi.org/10.1016/j.cub.2011.08.003.\u003c/li\u003e\n\u003cli\u003eGorgolewski, K., Burns, C.D., Madison, C., Clark, D., Halchenko, Y.O., Waskom, M.L. and Ghosh, S.S. (2011) \u0026lsquo;Nipype: A Flexible, Lightweight and Extensible Neuroimaging Data Processing Framework in Python\u0026rsquo;, \u003cem\u003eFrontiers in Neuroinformatics\u003c/em\u003e, 5. Available at: https://doi.org/10.3389/fninf.2011.00013.\u003c/li\u003e\n\u003cli\u003eGorgolewski, K.J., Auer, T., Calhoun, V.D., Craddock, R.C., Das, S., Duff, E.P., Flandin, G., Ghosh, S.S., Glatard, T., Halchenko, Y.O., Handwerker, D.A., Hanke, M., Keator, D., Li, X., Michael, Z., Maumet, C., Nichols, B.N., Nichols, T.E., Pellman, J., Poline, J.-B., Rokem, A., Schaefer, G., Sochat, V., Triplett, W., Turner, J.A., Varoquaux, G. and Poldrack, R.A. (2016) \u0026lsquo;The brain imaging data structure, a format for organizing and describing outputs of neuroimaging experiments\u0026rsquo;, \u003cem\u003eScientific Data\u003c/em\u003e, 3(1), p. 160044. Available at: https://doi.org/10.1038/sdata.2016.44.\u003c/li\u003e\n\u003cli\u003eHampson, M., Ruiz, S. and Ushiba, J. (2020) \u0026lsquo;Neurofeedback\u0026rsquo;, \u003cem\u003eNeuroImage\u003c/em\u003e, 218, p. 116473. Available at: https://doi.org/10.1016/j.neuroimage.2019.116473.\u003c/li\u003e\n\u003cli\u003eHayashi, M., Okuyama, K., Mizuguchi, N., Hirose, R., Okamoto, T., Kawakami, M. and Ushiba, J. (2022) \u0026lsquo;Spatially bivariate EEG-neurofeedback can manipulate interhemispheric inhibition\u0026rsquo;, \u003cem\u003eeLife\u003c/em\u003e, 11, p. e76411. Available at: https://doi.org/10.7554/eLife.76411.\u003c/li\u003e\n\u003cli\u003eJahanian, H., Holdsworth, S., Christen, T., Wu, H., Zhu, K., Kerr, A.B., Middione, M.J., Dougherty, R.F., Moseley, M. and Zaharchuk, G. (2019) \u0026lsquo;Advantages of short repetition time resting-state functional MRI enabled by simultaneous multi-slice imaging\u0026rsquo;, \u003cem\u003eJournal of Neuroscience Methods\u003c/em\u003e, 311, pp. 122\u0026ndash;132. Available at: https://doi.org/10.1016/j.jneumeth.2018.09.033.\u003c/li\u003e\n\u003cli\u003eKadosh, K.C. and Staunton, G. (2019) \u0026lsquo;A systematic review of the psychological factors that influence neurofeedback learning outcomes\u0026rsquo;, \u003cem\u003eNeuroImage\u003c/em\u003e, 185, pp. 545\u0026ndash;555. Available at: https://doi.org/10.1016/j.neuroimage.2018.10.021.\u003c/li\u003e\n\u003cli\u003eLi, X., Morgan, P.S., Ashburner, J., Smith, J. and Rorden, C. (2016) \u0026lsquo;The first step for neuroimaging data analysis: DICOM to NIfTI conversion\u0026rsquo;, \u003cem\u003eJournal of Neuroscience Methods\u003c/em\u003e, 264, pp. 47\u0026ndash;56. Available at: https://doi.org/10.1016/j.jneumeth.2016.03.001.\u003c/li\u003e\n\u003cli\u003eOldfield, R.C. (1971) \u0026lsquo;The assessment and analysis of handedness: The Edinburgh inventory\u0026rsquo;, \u003cem\u003eNeuropsychologia\u003c/em\u003e, 9(1), pp. 97\u0026ndash;113. Available at: https://doi.org/10.1016/0028-3932(71)90067-4.\u003c/li\u003e\n\u003cli\u003ePereira, J., Direito, B., Sayal, A., Ferreira, C. and Castelo-Branco, M. (2019) \u0026lsquo;Self-modulation of premotor cortex interhemispheric connectivity in a real-time functional magnetic resonance imaging neurofeedback study using an adaptive approach\u0026rsquo;, \u003cem\u003eBrain Connectivity\u003c/em\u003e, 9(9), pp. 662\u0026ndash;672.\u003c/li\u003e\n\u003cli\u003eRamot, M. and Martin, A. (2022) \u0026lsquo;Closed-loop neuromodulation for studying spontaneous activity and causality\u0026rsquo;, \u003cem\u003eTrends in Cognitive Sciences\u003c/em\u003e, 26(4), pp. 290\u0026ndash;299. Available at: https://doi.org/10.1016/j.tics.2022.01.008.\u003c/li\u003e\n\u003cli\u003eRenz, M.P., Zidda, F., Andoh, J., Prager, M., Sack, M., Becker, R., Ruf, M., Schmitgen, M.M., Wolf, R.C., Meyer‐Lindenberg, A. and Tost, H. (2023) \u0026lsquo;Practical challenges of continuous real‐time functional magnetic resonance imaging neurofeedback with multiband accelerated echo‐planar imaging and short repetition times\u0026rsquo;, \u003cem\u003eHuman Brain Mapping\u003c/em\u003e, 44(3), pp. 1278\u0026ndash;1282. Available at: https://doi.org/10.1002/hbm.26154.\u003c/li\u003e\n\u003cli\u003eRisk, B.B., Kociuba, M.C. and Rowe, D.B. (2018) \u0026lsquo;Impacts of simultaneous multislice acquisition on sensitivity and specificity in fMRI\u0026rsquo;, \u003cem\u003eNeuroImage\u003c/em\u003e, 172, pp. 538\u0026ndash;553. Available at: https://doi.org/10.1016/j.neuroimage.2018.01.078.\u003c/li\u003e\n\u003cli\u003eRos, T., Enriquez-Geppert, S., Zotev, V., Young, K.D., Wood, G., Whitfield-Gabrieli, S., Wan, F., Vuilleumier, P., Vialatte, F., Van De Ville, D., Todder, D., Surmeli, T., Sulzer, J.S., Strehl, U., Sterman, M.B., Steiner, N.J., Sorger, B., Soekadar, S.R., Sitaram, R., Sherlin, L.H., Sch\u0026ouml;nenberg, M., Scharnowski, F., Schabus, M., Rubia, K., Rosa, A., Reiner, M., Pineda, J.A., Paret, C., Ossadtchi, A., Nicholson, A.A., Nan, W., Minguez, J., Micoulaud-Franchi, J.-A., Mehler, D.M.A., L\u0026uuml;hrs, M., Lubar, J., Lotte, F., Linden, D.E.J., Lewis-Peacock, J.A., Lebedev, M.A., Lanius, R.A., K\u0026uuml;bler, A., Kranczioch, C., Koush, Y., Konicar, L., Kohl, S.H., Kober, S.E., Klados, M.A., Jeunet, C., Janssen, T.W.P., Huster, R.J., Hoedlmoser, K., Hirshberg, L.M., Heunis, S., Hendler, T., Hampson, M., Guggisberg, A.G., Guggenberger, R., Gruzelier, J.H., G\u0026ouml;bel, R.W., Gninenko, N., Gharabaghi, A., Frewen, P., Fovet, T., Fern\u0026aacute;ndez, T., Escolano, C., Ehlis, A.-C., Drechsler, R., Christopher deCharms, R., Debener, S., De Ridder, D., Davelaar, E.J., Congedo, M., Cavazza, M., Breteler, M.H.M., Brandeis, D., Bodurka, J., Birbaumer, N., Bazanova, O.M., Barth, B., Bamidis, P.D., Auer, T., Arns, M. and Thibault, R.T. (2020) \u0026lsquo;Consensus on the reporting and experimental design of clinical and cognitive-behavioural neurofeedback studies (CRED-nf checklist)\u0026rsquo;, \u003cem\u003eBrain\u003c/em\u003e, 143(6), pp. 1674\u0026ndash;1685. Available at: https://doi.org/10.1093/brain/awaa009.\u003c/li\u003e\n\u003cli\u003eRotem Botvinik-Nezer, Felix Holzmeister, Lei Zhang, Xu Zhang, Joshua E. Zosky, Thomas E. Nichols, Russell A. Poldrack, and Tom Schonberg (2020) \u0026lsquo;Variability in the analysis of a single neuroimaging dataset by many teams\u0026rsquo;, \u003cem\u003eNature\u003c/em\u003e [Preprint]. Available at: https://doi.org/10.1038/s41586-020-2314-9.\u003c/li\u003e\n\u003cli\u003eSayal, A., Sousa, T., Duarte, J.V., Costa, G.N., Martins, R. and Castelo-Branco, M. (2020) \u0026lsquo;Identification of competing neural mechanisms underlying positive and negative perceptual hysteresis in the human visual system\u0026rsquo;, \u003cem\u003eNeuroImage\u003c/em\u003e, 221, p. 117153.\u003c/li\u003e\n\u003cli\u003eScharnowski, F., Hutton, C., Josephs, O., Weiskopf, N. and Rees, G. (2012) \u0026lsquo;Improving Visual Perception through Neurofeedback\u0026rsquo;, \u003cem\u003eThe Journal of Neuroscience\u003c/em\u003e, 32(49), pp. 17830\u0026ndash;17841. Available at: https://doi.org/10.1523/JNEUROSCI.6334-11.2012.\u003c/li\u003e\n\u003cli\u003eSeth, A.K., Chorley, P. and Barnett, L.C. (2013) \u0026lsquo;Granger causality analysis of fMRI BOLD signals is invariant to hemodynamic convolution but not downsampling\u0026rsquo;, \u003cem\u003eNeuroImage\u003c/em\u003e, 65, pp. 540\u0026ndash;555. Available at: https://doi.org/10.1016/j.neuroimage.2012.09.049.\u003c/li\u003e\n\u003cli\u003eSitaram, R., Ros, T., Stoeckel, L., Haller, S., Scharnowski, F., Lewis-Peacock, J., Weiskopf, N., Blefari, M.L., Rana, M., Oblak, E., Birbaumer, N. and Sulzer, J. (2017) \u0026lsquo;Closed-loop brain training: the science of neurofeedback\u0026rsquo;, \u003cem\u003eNature Reviews Neuroscience\u003c/em\u003e, 18(2), pp. 86\u0026ndash;100. Available at: https://doi.org/10.1038/nrn.2016.164.\u003c/li\u003e\n\u003cli\u003eSorger, B., Kamp, T., Weiskopf, N., Peters, J.C. and Goebel, R. (2018) \u0026lsquo;When the Brain Takes \u0026ldquo;BOLD\u0026rdquo; Steps: Real-Time fMRI Neurofeedback Can Further Enhance the Ability to Gradually Self-regulate Regional Brain Activation\u0026rsquo;, \u003cem\u003eNeuroscience\u003c/em\u003e, 378, pp. 71\u0026ndash;88. Available at: https://doi.org/10.1016/j.neuroscience.2016.09.026.\u003c/li\u003e\n\u003cli\u003eSousa, T., Direito, B., Lima, J., Ferreira, C., Nunes, U. and Castelo-Branco, M. (2016) \u0026lsquo;Control of Brain Activity in hMT+/V5 at Three Response Levels Using fMRI-Based Neurofeedback/BCI\u0026rsquo;, \u003cem\u003ePLOS ONE\u003c/em\u003e. Edited by J. Bodurka, 11(5), p. e0155961. Available at: https://doi.org/10.1371/journal.pone.0155961.\u003c/li\u003e\n\u003cli\u003eSousa, T., Duarte, J.V., Costa, G.N., Kemper, V.G., Martins, R., Goebel, R. and Castelo-Branco, M. (2019) \u0026lsquo;Tracking perceptual decision mechanisms through changes in interhemispheric functional connectivity in human visual cortex\u0026rsquo;, \u003cem\u003eScientific Reports\u003c/em\u003e, 9(1), p. 1242. Available at: https://doi.org/10.1038/s41598-018-37822-x.\u003c/li\u003e\n\u003cli\u003eSousa, T., Sayal, A., Duarte, J.V., Costa, G.N. and Castelo-Branco, M. (2024) \u0026lsquo;A human cortical adaptive mutual inhibition circuit underlying competition for perceptual decision and repetition suppression reversal\u0026rsquo;, \u003cem\u003eNeuroImage\u003c/em\u003e, 285, p. 120488. Available at: https://doi.org/10.1016/j.neuroimage.2023.120488.\u003c/li\u003e\n\u003cli\u003eSousa, T., Sayal, A., Duarte, J.V., Costa, G.N., Martins, R. and Castelo-Branco, M. (2018) \u0026lsquo;Evidence for distinct levels of neural adaptation to both coherent and incoherently moving visual surfaces in visual area hMT+\u0026rsquo;, \u003cem\u003eNeuroImage\u003c/em\u003e, 179, pp. 540\u0026ndash;547. Available at: https://doi.org/10.1016/j.neuroimage.2018.06.075.\u003c/li\u003e\n\u003cli\u003eSrirangarajan, T., Mortazavi, L., Bortolini, T., Moll, J. and Knutson, B. (2021) \u0026lsquo;Multi‐band FMRI compromises detection of mesolimbic reward responses\u0026rsquo;, \u003cem\u003eNeuroImage\u003c/em\u003e, 244, p. 118617. Available at: https://doi.org/10.1016/j.neuroimage.2021.118617.\u003c/li\u003e\n\u003cli\u003eTodd, N., Josephs, O., Zeidman, P., Flandin, G., Moeller, S. and Weiskopf, N. (2017) \u0026lsquo;Functional Sensitivity of 2D Simultaneous Multi-Slice Echo-Planar Imaging: Effects of Acceleration on g-factor and Physiological Noise\u0026rsquo;, \u003cem\u003eFrontiers in Neuroscience\u003c/em\u003e, 11. Available at: https://doi.org/10.3389/fnins.2017.00158.\u003c/li\u003e\n\u003cli\u003eTodd, N., Moeller, S., Auerbach, E.J., Yacoub, E., Flandin, G. and Weiskopf, N. (2016) \u0026lsquo;Evaluation of 2D multiband EPI imaging for high-resolution, whole-brain, task-based fMRI studies at 3T: Sensitivity and slice leakage artifacts\u0026rsquo;, \u003cem\u003eNeuroImage\u003c/em\u003e, 124, pp. 32\u0026ndash;42. Available at: https://doi.org/10.1016/j.neuroimage.2015.08.056.\u003c/li\u003e\n\u003cli\u003eTyszka, M. (2023) \u0026lsquo;bidskit\u0026rsquo;. Available at: https://github.com/jmtyszka/bidskit.\u003c/li\u003e\n\u003cli\u003eWall, M.B. (2023) \u0026lsquo;Multiband acquisition sequences for fMRI: Proceed with caution\u0026rsquo;, \u003cem\u003eAperture Neuro\u003c/em\u003e, 3. Available at: https://doi.org/10.52294/001c.91292.\u003c/li\u003e\n\u003cli\u003eWang, T., Peeters, R., Mantini, D. and Gillebert, C.R. (2020) \u0026lsquo;Modulating the interhemispheric activity balance in the intraparietal sulcus using real-time fMRI neurofeedback: Development and proof-of-concept\u0026rsquo;, \u003cem\u003eNeuroImage: Clinical\u003c/em\u003e, 28, p. 102513. Available at: https://doi.org/10.1016/j.nicl.2020.102513.\u003c/li\u003e\n\u003cli\u003eWeber, L.A., Ethofer, T. and Ehlis, A.-C. (2020) \u0026lsquo;Predictors of neurofeedback training outcome: A systematic review\u0026rsquo;, \u003cem\u003eNeuroImage: Clinical\u003c/em\u003e, 27, p. 102301. Available at: https://doi.org/10.1016/j.nicl.2020.102301.\u003c/li\u003e\n\u003cli\u003eYarkoni, T., Poldrack, R.A., Nichols, T.E., Van Essen, D.C. and Wager, T.D. (2011) \u0026lsquo;Large-scale automated synthesis of human functional neuroimaging data\u0026rsquo;, \u003cem\u003eNature Methods\u003c/em\u003e, 8(8), pp. 665\u0026ndash;670. Available at: https://doi.org/10.1038/nmeth.1635.\u003c/li\u003e\n\u003cli\u003eZhang, Y., Ye, Z., Xia, C., Tan, Y., Zhang, M., Lv, X., Tang, J. and Li, Z. (2024) \u0026lsquo;Clinical Applications and Recent Updates of Simultaneous Multi-slice Technique in Accelerated MRI\u0026rsquo;, \u003cem\u003eAcademic Radiology\u003c/em\u003e, p. S1076633223007109. Available at: https://doi.org/10.1016/j.acra.2023.12.032.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Visual perception, bistable motion, interhemispheric connectivity, neurofeedback, multiband, fMRI","lastPublishedDoi":"10.21203/rs.3.rs-4335511/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4335511/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eFunctional magnetic resonance imaging (fMRI) allows to observe neural activity in real-time but tracking the neural correlates of perceptual decision as a function of interhemispheric connectivity has remained difficult. Recent advances in image acquisition, namely with the surfacing of multiband sequences, have led us to investigate this mechanism using higher temporal resolution approaches. We were able to better capture the hemodynamic responses to rapid changes in neural activity concomitantly with a task requiring either perceptual interhemispheric segregation or integration, shortening the gap to other neuroimaging techniques, which is particularly significant when considering the study of dynamic connectivity patterns. Here, we tested the hypothesis whether interhemispheric connectivity in the visual cortex relates to interhemispheric integration, when presented with bistable moving stimuli at four distinct temporal resolutions. Based on this connectivity metric, we could discern perceptual state transitions related to connectivity. First, we found that activation response metrics to visual motion in our target region of interest, the human visual motion complex hMT+, are stable across temporal resolutions. Then, we investigated interhemispheric connectivity between homologous hMT\u0026thinsp;+\u0026thinsp;in response to bistable moving stimuli, for all resolutions, which was critical for replication of perception related interhemispheric synchrony. The established relation between perceptual coherence and increased synchrony across the hemispheres suggests the feasibility of a real-time fMRI neurofeedback based on interhemispheric connectivity. Accordingly, we could infer perceptual states based on this connectivity metric while designing a rule that could even be used to generate feedback. We further showed that higher resolution sequences are beneficial when implementing feedback interfaces based on interhemispheric functional connectivity, both regarding the delay and the accuracy of the feedback itself. Regarding the use of real time fMRI and neurofeedback strategies, higher resolution sequences are likely needed, when relying on connectivity metrics.\u003c/p\u003e","manuscriptTitle":"Interhemispheric connectivity during spontaneous switches in visual perceptual coherence as revealed by fMRI at multiple temporal resolutions","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-05-08 20:29:12","doi":"10.21203/rs.3.rs-4335511/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"dad37f99-88df-4496-a920-169eeb66879f","owner":[],"postedDate":"May 8th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":31461472,"name":"Biological sciences/Neuroscience/Cognitive neuroscience/Perception"},{"id":31461473,"name":"Biological sciences/Neuroscience/Sensory processing"},{"id":31461474,"name":"Biological sciences/Neuroscience/Neural circuit"}],"tags":[],"updatedAt":"2024-09-25T18:38:30+00:00","versionOfRecord":[],"versionCreatedAt":"2024-05-08 20:29:12","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4335511","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4335511","identity":"rs-4335511","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}