THINC: A Hybrid Methodological Framework for Human-Computer Interaction Analysis – A Case Study of Design Thinking within Virtual Reality

THINC: A Hybrid Methodological Framework for Human-Computer Interaction Analysis – A Case Study of Design Thinking within Virtual Reality | Authorea try { document.documentElement.classList.add('js'); } catch (e) { } var _gaq = _gaq || []; _gaq.push(['_setAccount', 'G-8VDV14Y67G']); _gaq.push(['_trackPageview']); (function() { var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true; ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js'; var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s); })(); Skip to main content Preprints Collections Wiley Open Research IET Open Research Ecological Society of Japan All Collections About About Authorea FAQs Contact Us Quick Search anywhere Search for preprint articles, keywords, etc. Search Search ADVANCED SEARCH SCROLL This is a preprint and has not been peer reviewed. Data may be preliminary. 8 January 2026 V1 Latest version Share on THINC: A Hybrid Methodological Framework for Human-Computer Interaction Analysis – A Case Study of Design Thinking within Virtual Reality Authors : Soroush Masoumzadeh 0000-0003-3295-6616 [email protected] , Rongrong Yu , Ina bornkessel-schlesewsky , Ning Gu , Fan Zhang , Jimmy Cao , Ben Volmer , and Adam Drogemuller Authors Info & Affiliations https://doi.org/10.22541/au.176790582.28137799/v1 198 views 97 downloads Contents Abstract INTRODUCTION A REVIEW OF EXISTING METHODS IN HCI THINC: A HYBRID METHODOLOGICAL FRAMEWORK CASE STUDY: THE IMPACT OF LEVEL OF FIDELITY (LOF) ON DESIGN THINKING IN IVE Data Analysis Results Discussion CONCLUSION References Information & Authors Metrics & Citations View Options References Figures Tables Media Share Abstract This chapter introduces a methodological framework we call THINC—Tracking Human-computer Interaction through Neural and Cognitive data—which is designed to understand human-computer interactions (HCI) by integrating neural and cognitive data. THINC serves as a framework that tries to connect objective and subjective data, combining them to provide a new perspective and deeper insights into human interactions with computers, especially the recently developed immersive technologies. We first explore existing approaches, highlighting their strengths and limitations, and then propose THINC as a hybrid solution that attempts to integrate objective and subjective data to gain deeper, real-time insights. To illustrate its application, we present and discuss a design thinking study conducted within immersive virtual environments (IVE) with different Levels of Fidelity (LoF) to see how the LoF might influence design thinking. Through this case study, the chapter demonstrates how THINC allows us to better understand the differences between cognitive processes of designers as they engage with immersive environments with different levels of realism. Keywords: Neurocognitive analysis, Human-computer Interaction (HCI), Design thinking, Virtual Reality, Protocol analysis, Electroencephalography (EEG), Design cognition, Design neurocognition, Methodology, Fidelity, Mixed methods Soroush Masoumzadeh 1 , Rongrong Yu 1 , Ina Bornkessel-Schlesewsky 1 , Ning Gu 1 , Fan Zhang 2 , Jimmy Cao 1 , Ben Volmer 1 , Adam Drogemuller 1 1 University of South Australia, Adelaide, South Australia 5000, Australia 2 Griffith University. Gold Coast, Queensland, 4222, Australia Keywords: Neurocognitive analysis, Human-computer Interaction (HCI), Design thinking, Virtual Reality, Protocol analysis, Electroencephalography (EEG), Design cognition, Design neurocognition, Methodology, Fidelity, Mixed methods INTRODUCTION HCI is typically defined as the interdisciplinary study of how people engage with computing systems, encompassing aspects of design, technology, and human factors (Dix et al. 2004, Carroll 2013). It integrates principles from computer science, psychology, cognitive science, and design to develop user-centred technologies that improve usability and user experience (Zhang and Adipat 2005, Lazar et al. 2017). As digital interactions become increasingly complex, understanding how users think, behave, and respond to interfaces becomes critical for creating effective and intuitive systems (Norman 2013). For example, research on immersive environments has shown how augmented and virtual reality tools provide new potentials for user interaction by more deeply integrating software and hardware design principles to support embodied, spatial, and multisensory experiences (Thomas & Piekarski, 2002; Maher et al., 2011). More recent work in creativity support tools and natural user interfaces has further highlighted the need for evaluating how such systems influence users’ cognitive states, affective responses, and behaviour in rich, multi-sensory settings (Remy et al., 2020; Jain et al., 2011). Various methods have been developed for studying HCI, each providing insights into user interaction. Empirical approaches, such as usability testing and controlled experiments, generate data on system performance and user behaviour (Lazar et al., 2017; Gómez-Zará et al., 2024). These methods help evaluate interface efficiency, identify usability issues, and inform iterative design improvements. Likewise, field studies and ethnographic research document user behaviour in real-world settings, highlighting the contextual factors influencing interaction (Rogers et al., 2011; Crabtree et al., 2003). Other methods, including think-aloud protocols and protocol analysis, have been widely employed to explore user perceptions and cognitive processes during interaction (Bargas-Avila and Hornbæk, 2011; Frich et al., 2019; Ericsson and Simon, 1993). These approaches uncover subjective experiences and decision-making patterns that may be hidden in quantitative data alone. Furthermore, cognitive modeling simulates human thought processes to predict user behaviour (Kieras and Meyer, 2020; Newell and Card, 1985). Researchers in design cognition suggest that combining controlled experiments with qualitative, context-rich data can offer a more complete view of how users and designers process information within complex environments (Maher et al. 2011, Lockton et al. 2010). Despite their strengths, existing HCI research methods have limitations. Traditional usability testing and behavioural observations provide valuable insights into what users do but direct measures of brain activity may be able to provide insights into early processing stages that precede conscious decision-making, thus complementing subjective data; Examples include Electroencephalography (EEG) responses related to performance monitoring (e.g. see Ullsperger et al. 2014). Qualitative methods, while rich in contextual depth and essential for understanding user experience, are sometimes critiqued for their perceived subjectivity and limited generalisability (Blandford et al. 2016, Dourish 2006). However, these methods remain critical for capturing the nuance and meaning behind user actions, especially in design-focused research. On the other hand, cognitive modelling and neural methods—such as EEG—can measure users’ patterns of brain activity and physiological responses, but they frequently lack sufficient contextualization in relation to users’ explicit thoughts and intentions (Luck, 2014, Cowley et al. 2016). The divide between subjective and objective approaches partly results from the absence of a robust framework for their integration, leading many studies to prioritize one perspective at the expense of the other. Bridging this gap could yield a more holistic understanding of user experiences by connecting explicit cognitive processes to measurable neural and behavioural data. In an attempt to address these challenges, this chapter introduces a hybrid methodology that integrates subjective coding schemes—such as protocol analysis—with neuroscientific methods like EEG. By combining qualitative cognitive data with objective neural and physiological signals, we aim to offer a more comprehensive approach to understanding how users think about and experience digital interactions. In our case study examining design thinking within an IVE, we demonstrate the application of our proposed methodology by employing an adjusted version of the widely used Function-Behaviour-Structure (FBS) coding scheme (Gero and Kannengiesser 2004) in conjunction with EEG. This approach allows us to quantitatively and qualitatively explore how designers navigate through different stages of design within different kinds of VR interfaces. Beyond identifying patterns in user behaviour, our proposed method highlights the relationship between explicit verbalizations and underlying neural processes—particularly beneficial for understanding complex tasks such as design ideation (Maher et al. 2011, Lockton et al. 2010). A REVIEW OF EXISTING METHODS IN HCI HCI research has evolved over recent decades, incorporating diverse methods to better understand how users engage with digital systems. As computational systems grow increasingly complex, it has become essential to understand not only what users do but also how they think and experience these interactions (Rogers et al. 2011, Norman 2013). The emergence of immersive technologies—such as VR and AR—has particularly transformed HCI by enabling interactions that are more natural, embodied, and multisensory (Bailenson 2018). These technologies allow users to engage with digital environments beyond traditional screen-based interfaces, utilising spatial navigation, gesture-based controls, and even brain-computer interfaces (Zhang et al. 2025) Consequently, existing HCI research methods would benefit from the ability to effectively capture the cognitive processes underlying these interactions. In this section, we review the key methodological approaches in HCI, highlighting their contributions and limitations, and lay the groundwork for introducing the THINC method. Empirical research methods in HCI primarily evaluate system performance and user behaviour through controlled experiments, usability testing, and field studies (Dix et al. 2004, Crabtree et al. 2003). Usability testing remains foundational in HCI research, offering structured assessments of interface effectiveness and user satisfaction. Controlled experiments, usually conducted in laboratory settings, allow precise measurement of task completion times, error rates, and efficiency (Zhang et al. 2025). Although these methods generate robust data on user interactions, they mostly capture observable behaviours and rarely provide deeper insights into cognitive processes and decision-making (Lazar et al. 2017). Field studies and ethnographic research address this gap by studying user behaviour in real-world environments (Rogers et al. 2011). These approaches emphasize the contextual factors shaping interactions, including social dynamics, environmental conditions, and task-specific constraints (Crabtree et al. 2003). However, while strong in ecological validity, these methods often lack control over experimental variables, complicating efforts to isolate specific cognitive or neural mechanisms underlying user behaviour (Aldayel et al. 2020). For instance, a study comparing various VR input devices for older adults evaluated user performance and preferences across different interaction modalities (Healy et al. 2022). Additionally, a systematic review by Gómez-Zará et al. (2024) examined empirical studies on VR’s use for social interactions, highlighting how controlled experiments and observational methods have demonstrated VR’s effectiveness in facilitating social connections. Researchers have long used cognitive approaches, such as protocol analysis and cognitive modelling, to uncover the thought processes underpinning HCI (Newell and Card 1985, Kieras and Meyer 2020). Protocol analysis, where participants verbalize thoughts while performing tasks, has extensively informed research on design, problem-solving, and interface interactions (Bargas-Avila and Hornbæk 2011, Ericsson and Simon 1993). This approach helps researchers identify cognitive strategies, decision-making patterns, and user conceptualizations of digital interfaces. One widely used framework for protocol analysis is the FBS ontology. Gero and Kannengiesser (2014) utilized the FBS framework to examine design cognition, illustrating how designers dynamically transition between function, behaviour, and structure during design tasks. Their extensive work demonstrated the effectiveness of FBS in capturing the iterative nature of cognitive activities. Other cognitive modelling methods, like the Goals, Operators, Methods, and Selection rules (GOMS) model (Card et al. 1983), predict user performance by decomposing tasks into hierarchical steps, indicating that interfaces requiring fewer operational steps typically enhance efficiency. Additionally, Cognitive Load Theory (Sweller 1988) has informed digital learning environment designs, with evidence suggesting that simplified layouts reduce cognitive overload and enhance user engagement. Researchers have also developed structured coding schemes to systematically analyze interactions. For example, the ACT4Teams coding scheme (Kauffeld et al. 2018) evaluates communication, coordination, and decision-making in collaborative software development, while the ConcurTaskTrees (CTT) notation (Paternò 2004) models user tasks hierarchically, allowing detailed analysis of task sequences and dependencies. Despite their strengths, cognitive methods like protocol analysis have limitations. Protocol analysis relies heavily on self-reporting, introducing potential biases and neglecting unconscious cognitive processes. Cognitive models also often rely on simplified assumptions, potentially missing complexities inherent in human cognition. Combining these diverse cognitive approaches and structured coding schemes can provide a richer analysis, allowing researchers to triangulate subjective verbalizations with structured interaction data for deeper insights into user behaviour. Neuroscientific techniques, such as EEG, functional Near-Infrared Spectroscopy (fNIRS), and Magnetoencephalography (MEG), are increasingly applied in HCI research to objectively measure cognitive and emotional states (Luck 2014, Cowley et al. 2016). EEG, for instance, measures the electrical activity of the brain. One EEG-based analysis approach isolates event-related potentials (ERPs)—small brain signals triggered by specific events or stimuli – thus allowing researchers to evaluate cognitive workload (the mental effort required by a task), engagement (how involved or interested users are), and preference and attention levels in real time (Zhang et al. 2025). EEG activity can also be analysed in the frequency domain, examining oscillations in brainwave frequencies such as alpha: ~8-13 Hz (often associated with selective attention, inhibition of irrelevant stimuli, and creative ideation, often increasing when individuals engage in abstract, visual, or internally focused thinking (Klimesch 1999, Fink and Benedek, 2014; Benedek et al. 2014)). In design contexts, alpha desynchronization (reduction of alpha power) has been linked to visual attention and problem framing, especially in early conceptual stages (Liang et al. 2017, Liu et al. 2018), beta: ~13-30 Hz (Elevated beta oscillation is often seen during phases of design evaluation or constraint-driven reasoning (Liu et al. 2019), while reduced beta has been noted during divergent or intuitive stages of idea generation (Chaoyun et al. 2020), and theta waves: ~4-7 Hz (linked to memory encoding and reflective processing, have been consistently associated with moments of insight, concept restructuring, and cognitive flexibility in both creativity and design EEG studies (Mitchell et al. 2008, Jauk et al. 2012 Vieira et al. 2019, Gero and Milovanovic 2020). For instance, theta synchronization (increase in theta power) has been observed when expert designers reinterpret constraints or shift between design frames (Li et al. 2021, Lou et al. 2020). Monitoring these frequency bands helps researchers understand changes in mental effort, attention, and cognitive states during tasks (Makeig et al. 2004). Although fNIRS provides lower temporal resolution—meaning it captures changes in brain activity at slower timescales—it captures information about blood-flow changes (hemodynamic responses) and has a better spatial resolution than scalp-recorded EEG (i.e., it can shed light on which brain regions are activated during a particular task). For example, fNIRS data can reveal activation changes in the prefrontal cortex, a part of the brain involved in decision-making and higher cognitive functions, complementing EEG data in HCI studies (Pinti et al. 2020). MEG, which records magnetic fields generated by brain activity, combines higher spatial resolution than offered by EEG with high temporal resolution (tracking rapid brain changes), making it especially useful for studies on sensory processing, that is how the brain interprets sensory information, and detailed neural dynamics during interactions with digital environments, due to its comparatively superior ability to localize brain activity sources compared to EEG, although it is still not as precise as fMRI. (Baillet 2017). Several studies highlight the strengths of neuroscientific methods in HCI. For instance, research using mobile EEG demonstrated correlations between alpha wave changes and spatial awareness in virtual environments. Recent work by Aldayel et al. (2020) used EEG signals to classify user preferences, achieving above-chance accuracy in predicting favoured AR input methods. Their results highlight the potential of neuroadaptive systems to personalize interactions by decoding neurocognitive responses. Similarly, fNIRS has been applied to study cognitive load in adaptive learning environments, showing increased prefrontal activation with greater task difficulty (Ayaz and Dehais, 2019). These examples illustrate the effectiveness of neuroscientific tools in capturing cognitive and affective states in real time during user interactions. However, neuroscientific methods face challenges regarding ecological validity, data interpretation, and integration with qualitative methods (Cowley et al. 2016). Although neural data offer objective insights, they often lack context explaining why certain responses occur, requiring complementary qualitative approaches. Remy et al. (2020) and Frich et al. (2019) suggest that bridging this gap requires multi-method designs that combine physiological signals with user interviews, task performance metrics, and long-term observational data. Integrating neuroscientific with cognitive and behavioural approaches thus has the potential to provide a more comprehensive understanding of user experience, cognition, and emotional responses. The literature review highlights a potential lack of integration between subjective and objective methodologies in HCI research. Traditional behavioural and cognitive methods, such as protocol analysis and think-aloud protocols, provide detailed qualitative insights but lack the precision and continuous measurement capability of neuroscientific methods. Conversely, neuroscientific methods like EEG generate robust quantitative data about cognitive and physiological states but often fail to capture nuanced, context-dependent decision-making processes underlying interactions (Dourish 2006). This methodological divide limits comprehensiveness and ecological validity in HCI research, underscoring the need for integrated approaches that effectively combine qualitative cognitive data with precise physiological measures. Although some studies have begun integrating self-report and neurophysiological data, a broadly applicable, structured methodological framework is still lacking. A hybrid framework systematically integrating subjective cognitive categorizations with objective physiological measures would enhance interpretability, allowing researchers to more fully triangulate user experiences. THINC: A HYBRID METHODOLOGICAL FRAMEWORK In response to these methodological challenges, we propose THINC (Tracking Human-computer Interaction through Neural and Cognitive data) as a generalized and flexible framework that integrates subjective cognitive categorization methods—such as protocol analysis, think-aloud protocols, and collaborative coding schemes—with neuroscientific and physiological techniques, including EEG. This integration enables a continuous, dynamic, real-time analysis of cognitive processes during human-computer interactions, effectively bridging qualitative and quantitative research methodologies. THINC consists of three interconnected methodological pillars (Fig. 1), each contributing essential components to the overall framework: 1) Subjective Cognitive Data Categorization : This pillar involves systematically coding user activities through structured cognitive frameworks such as FBS, GOMS, or cognitive load phases. Using structured categorization facilitates clear segmentation and meaningful interpretation of cognitive activities, providing explicit qualitative reference points for analysis. 2) Temporal Segmentation: Alignment of cognitive categorizations with physiological data is ensured through clearly defined temporal epochs—such as event-based, continuous, or second-by-second intervals. This approach allows direct correlation of specific cognitive activities with corresponding neural and physiological responses, enhancing temporal accuracy and interpretability. 3) Objective Data Integration: Neurophysiological signals captured through different methods, such as EEG, are complemented with additional physiological measures like eye-tracking, heart rate, and behavioural data. This multimodal integration supports robust triangulation and validation of cognitive interpretations, yielding richer insights into user cognition, emotional engagement, and decision-making during digital interactions. By applying THINC, researchers can address the common limitations in HCI studies, notably issues regarding ecological validity and interpretability associated with relying solely on qualitative or quantitative methods. The explicit integration of structured cognitive categorizations with physiological signals allows nuanced analyses of user cognition, informing actionable insights for designing user-centered technologies and interactive environments. Figure 1. The THINC methodological framework. A key strength of THINC lies in its adaptability across diverse HCI research domains. For example, in design research, THINC has the potential to enable analysis of cognitive state transitions by aligning FBS-based subjective categorizations with EEG data, revealing cognitive patterns related to creativity, problem-solving, and concept generation. Similarly, in UX research, THINC could effectively identify cognitive and emotional states—such as engagement, frustration, or cognitive overload; by combining emotion-based subjective coding schemes with objective physiological measurements. Through ongoing refinement and development, THINC can provide a new way of looking at how cognitive processes are conceptualized, studied, and interpreted in HCI, facilitating precise, holistic, and contextually rich understandings of user experiences and interactions in digital environments. Having introduced the THINC framework and its methodological advantages, we can now demonstrate the nuances of its practical application through our early findings of a case study focused on design thinking in IVEs. Our design thinking study represents an ideal context for illustrating THINC’s capabilities, given its inherent complexity, iterative nature, and the interplay between cognitive processes and contextual influences. By systematically integrating subjective cognitive categorizations—captured through verbal reflections—with objective neurophysiological data obtained from EEG recordings, this case study showcases THINC’s potential to provide new insights into how different types of virtual environments with varying LoFs influence design thinking. CASE STUDY: THE IMPACT OF LEVEL OF FIDELITY (LOF) ON DESIGN THINKING IN IVE Design problems are typically characterized as complex and ill-defined, involving iterative interactions between designers and evolving solutions. These iterative interactions often lead designers to redefine or reframe problems, resulting in innovative outcomes that emerge from continuous negotiations between problem and solution spaces. Even initially clear design tasks frequently become ambiguous as designers explore possibilities, generating novel ideas and alternative solutions. Despite its widespread use, design thinking still lacks a universally accepted definition; various studies conceptualize it as either a cognitive process or a product of specific disciplinary activities (Lee et al. 2020). In the following, we describe a subset of findings from a recent study (Masoumzadeh et al. 2025a) as a case study for application of the THINC framework. In this study, we adopt Gero’s (1990) conceptualization of design thinking as a cognitive activity involving the iterative co-evolution of problem and solution spaces, highlighting their mutual influence throughout the design process. This iterative relationship encourages designers to continually redefine problems, promoting deeper cognitive exploration and more innovative outcomes. In this research, we employed the THINC methodological framework to gain deeper insights into design cognition by integrating subjective cognitive data with objective neurophysiological responses. This approach bridges the interpretive limitations of qualitative methods and the contextual ambiguity of solely neuroscientific data. The objectives of this research are captured by the following research question: RQ1: Can patterns of EEG activity within specific frequency bands (i.e., alpha) be reliably predicted based on the design space (Problem space vs. Solution space) engaged by architects during concept design tasks? RQ2: How do High Fidelity (HF) and Low Fidelity (LF) virtual environments influence designers’ EEG patterns during architectural concept design? Research Design The experimental setup included a Meta Quest 3 VR headset, selected for its graphical capabilities, user-friendly interface, and ergonomic design, providing participants with an immersive yet comfortable environment for architectural concept design tasks. Gravity Sketch software was used for design activities due to its intuitive, organic interactions and minimal training requirements, making it ideal for concept generation in architectural design. Task performance and screen recordings were managed on a separate laptop equipped with OBS Studio, capturing records of participants’ interactions within the virtual environments. For EEG data collection, we used a wireless 16-channel LiveAmp system (Brain Products GmbH, Gilching, Germany), chosen for its portability, high signal quality, and minimal interference with the VR headset, ensuring accurate and reliable data collection ( see Masoumzadeh et al. 2025a for details of EEG data recording ). The experimental room was maintained as a controlled environment, minimising external noise and distractions. EEG data monitoring was conducted on a separate recording computer, while task performance and screen recording were handled by a laptop with sufficient GPU capabilities. The Meta Quest 3 headset was connected directly to the laptop using the Meta Quest Link Cable, leveraging the laptop’s GPU instead of the headset’s, to ensure smooth performance and avoid battery issues during tasks. Fig. 2 provides pictures of the setup. Figure 2. The setup of the experiment room and participants engaged in the design tasks. Participants (N=11, 4 females) included architecture practitioners with a minimum of two years professional experience, and PhD and Master’s students in architecture. These participants are a subset of a larger sample reported in Masoumzadeh et al. (2025a.) Participants’ ages ranged between 19–39 years, who were instructed to avoid caffeine, medication, or any substances that might influence cognitive performance prior to the experiment. Handedness was not considered as an inclusion criterion (Willems et al 2022). None of the participants reported hearing impairments, although two participants wore glasses. For these participants, the VR headset was adjusted to ensure comfort and usability during the experiment. Participants performed two architectural design tasks within VR, involving conceptual design ideation for two distinct architectural scenarios: a pop-up store in a religious neighbourhood (R) and a cafeteria in a modern neighbourhood (M). Both scenarios were explored under two levels of environmental fidelity—high-fidelity (H) and low-fidelity (L)—to examine how the fidelity of virtual environments influences design cognition (Teimouri et al. 2024). Task sequences varied systematically among participants (RL-MH—that is, Religious/Low-fidelity-Modern/High-fidelity , RH-ML, ML-RH, MH-RL) to minimize order effects and potential learning biases (Fig. 3). Additionally, participants completed two preliminary control tasks with high and low fidelity before engaging in the primary tasks. These included simplified and shorter (5-minute) versions of the 20-minute architectural design task. These control tasks served as a test environment to familiarize the participants with the new virtual environments before starting the main 20-minute task. Participants also completed a 2-minute resting-state task with eyes open and eyes closed, at the beginning and end of the experiment session, which provided individual alpha frequency (IAF), the peak frequency within the alpha rhythm. Recording IAF enables an individualized adjustment of frequency bands by accounting for inter-individual differences in alpha peak frequency, leading to more accurate definition of frequency bands (Kilmesch 1999). Figure 3. The different environments and LoFs employed in the experiment: a) Religious, Low-fidelity (RL), b) Religious, High-fidelity (RH), c) Modern, Low-fidelity (ML), and d) Modern, High-fidelity (MH). Participants were asked to design in the designated area shown in pink. In the LF environments (a & c) the function of the buildings was showcased using a color scheme—yellow for residential, red for commercial, and blue for religious. Each design task lasted 20 minutes , with a 5-minute break between tasks to minimize fatigue and reduce the likelihood of motion sickness (Fig. 4). Participants had freedom to explore and interact with the virtual environments; however, to accommodate the EEG setup and minimize movement-related noise in the EEG data, they remained seated during the design tasks and navigated the environment using handheld controllers. To ensure EEG data quality, participants were advised not to speak during tasks; however, brief communication with the researcher was permitted if issues arose or assistance was needed. EEG data recorded during such interactions were excluded from subsequent analysis. It is noteworthy that the procedure was approved by University of South Australia’s Ethics Committee (205884). After completing the VR tasks, participants reviewed recordings of their VR sessions and provided retrospective verbal accounts detailing their cognitive processes and decision-making. These verbal reflections were transcribed and analysed using the Function-Behaviour-Structure (FBS+C) coding scheme (Masoumzadeh et al. 2024), an extension of the original FBS ontology (Gero and Kannengiesser 2004). The FBS+C coding scheme categorizes design activities into: Function (F): referring to the intended purposes of designs, Behaviour (B): concerning the actions and performance of designs, Structure (S): referring to physical and structural components of designs. Context (C) and Behaviour of Context (Bc) categories were specifically included to capture contextual influences. The C code refers to external influences such as environmental conditions, user requirements, or situational constraints that potentially shape or affect the design process. The Bc code captures how design decisions or structural changes influence the surrounding context, potentially leading to new contextual considerations or modifications. In other words, C addresses how context influences the designer, whereas Bc shows how the designer’s actions, in turn, affect the context which is considered in the design process. While C and Bc can be further utilized and analysed to understand the influence of context on the design process, in this study they are coded and included specifically because the LoF of the virtual environments in which designers are engaged is measured; LoF is considered part of context and is partially captured within the C and Bc codes. Figure 4. The experiment stages and steps. Data Analysis To follow the first step of the THINC framework (Fig. 1), namely cognitive data categorization , the verbal data collected from participants were transcribed using Whisper.io, an automatic speech recognition tool developed by OpenAI, which converts audio recordings into highly accurate text transcripts. The transcripts were then reviewed by a researcher to ensure accuracy. Cognitive data were categorized using the FBS+C ontology, which was identified through designers’ Actions (by observing their recorded design activities) and verbal Statements (via retrospective interviews about their design activities), as outlined in specific columns in Table 1. The FBS+C coding scheme was then applied to this data across two temporal scales: the EEG column codes activities in two-second epochs as they occur during the design task, while the FBS+C column identifies the type of activities (i.e., design spaces) as designers move between different stages of the design process. Following this stage, the second step in the proposed methodology is temporal segmentation ; that is, preparing and categorising the cognitive data so that it can be interpreted in relation to how these processes unfold over specific time frames. For this, protocol analysis was specifically adapted (Table 1) for alignment with EEG analysis by coding the data in two-second epochs. The decision to use two-second epochs aimed to enhance coding feasibility, given the complexity of the study design involving multiple control and design tasks, as well as the total number of participants recruited in the final stage of the study, a subset of whom is reported in this case study. Table 1 An example table format for coding subjective protocol data in a way that is suitable for EEG analysis. 7 MH 00:00 *reads the task* So, here I’m reading the task to make sure I understand what I should do R R 7 MH 00:02 *reads the task* R 7 MH 00:04 *reads the task* R 7 MH 00:06 *zooms out* And then trying to see where I’m located at compared to the mosque C C 7 MH 00:08 *looks around* C 7 MH 00:10 *looks around* C 7 MH 00:12 *draws object* I started sketching with a simple shape, like a sphere S S 7 MH 00:14 *draws object* S 7 MH 00:15 *draws object* S 7 MH 00:16 *deletes object* But then realized it’s too avangard for a local cafe Bs Bs … … … … … … … As demonstrated in Table 1 , coding was based primarily on two types of data derived from the design sessions: ”Actions” and ”Statements.” Actions included observable activities in the virtual environment (VE) performed by participants, such as ”reads the task,” ”zooms out,” ”zooms in,” ”looks around,” ”looks at design,” ”navigates environment,” ”draws object,” ”deletes object,” ”moves object,” ”adjusting object,” ”browsing menu,” and ”copies object.” Each action corresponds to a particular cognitive activity influencing the design process. For instance, ”looks around” could reflect consideration of the Function (F) of the design or the collection of Contextual (C) information. After coding these actions into two-second epochs (see Temporal segmentation column in Table 1), researchers further reviewed transcripts of participants’ statements to gain deeper insights into their thought processes and decision-making—commonly referred to as the ”black box” of design thinking. These more extended segments, coded using the FBS+C framework, spanned from the beginning to the end of each coherent action or statement (see Design space protocols column in Table 1). Aligning these broader FBS+C-coded segments with EEG data enabled researchers to analyze correlations between specific cognitive activities and corresponding brain activities, offering deeper insights into the neurocognitive changes underpinning the design process. According to the THINC framework, the next step is neural data integration . The aim of this step is to process the raw neural data in a way that can be used alongside the qualitative data collected in the two previous steps. This begins with an EEG analysis pipeline, where the raw EEG signals are first pre-processed to ensure data quality and consistency. Pre-processing typically includes filtering to remove noise and unwanted frequency components, re-referencing the data (for example, to mastoid electrodes), and detecting and removing artefacts such as eye blinks and muscle activity—often using techniques like Independent Component Analysis (ICA). Noisy or faulty channels are also identified and excluded as needed ( see Masoumzadeh et al. 2025a and 2025b for details regarding EEG (pre)processing methods ). To enable the integration of processed EEG data with the behavioural and cognitive activities categorized in Table 1, it was necessary to segment the continuous EEG recordings into fixed two-second epochs. In this study, the start and end of each design session were marked by manual triggers set by the researcher using an EEG tracking application developed in Unity 2022.3.10f1. These triggers provided precise event markers, allowing the identification of task-relevant data segments for each participant and task. The EEG data were cropped to include only the design task period, with an additional two seconds appended at the beginning to allow for scaling relative to a pre-task period in subsequent analyses. The cropped segments were then divided into consecutive, non-overlapping two-second epochs, providing a consistent temporal framework for aligning neural signals with the behavioural coding scheme. For each recording, metadata—including both epoch labels in the Temporal segmentation column and cognitive activity codes derived from the Design space protocols column—were imported from protocol files and matched to the corresponding epochs (Fig. 5). Artefact rejection was carried out using the AutoReject algorithm (Jas et al., 2017), which automatically detects and removes epochs containing excessive noise or signal artefacts. Diagnostic plots were generated at this stage for quality assurance. Only clean, high-quality epochs were retained for subsequent analyses, and both the processed epochs and their associated metadata were saved to enable multimodal integration with the behavioural and cognitive datasets. Figure 5. A schematic visual showcasing the final dataset constructed by overlaying the outputs of each step of THINC. Results Linear mixed-effects models (LMMs) were computed with the lme4 package in R version 4.5.0 (Bates et al. 2015) to examine the effects of cognitive design stage (Problem space vs Solution space indicated by the F, Be, Bs, S, C, Bc coding scheme) and virtual environment fidelity (HF vs LF) on EEG power in the alpha band. While beta and theta bands was also analysed, the results for those bands are not reported here due to the focus of this case study on showcasing THINC ( see Masoumzadeh et al. 2025b for a full report of alpha, beta and theta bands ). Gamma band was omitted from these analyses, in line with prior recommendations regarding susceptibility to artifact above ∼30 Hz (Cross et al., 2018; Buzsáki and Schomburg 2015). RQ1: EEG Power by Design space (Problem vs Solution space) Analysis of the linear mixed-effects models revealed that mean predicted EEG power varied between problem and solution spaces across all frequency bands, but these differences were generally subtle. In the alpha band (Fig. 6), predicted power was higher in the problem space compared to the solution space, with the largest differences observed in central and anterior regions. In contrast, occipital regions showed slightly higher alpha power for the solution space, while the posterior regions demonstrated minimal mean differences. For the beta band (Fig. 8), mean predicted power was slightly higher for the problem space in anterior and central regions, with the remaining regions showing little difference between problem and solution spaces. Over time, beta power in central regions was consistently higher in the problem space, and a slight increase over epochs was seen in anterior regions for the problem space, while the left posterior remained parallel and the right posterior showed a gradual increase. Figure 6. Model-predicted EEG power by cognitive design space and ROI (left). Mean predicted power (86% CI) for alpha band across problem and solution spaces, separated by ROI. Predicted power over epochs (time) for alpha band, by ROI and cognitive space (right). In the theta band (Fig. 7), mean power for the problem space was slightly higher in central and left posterior regions and most visibly higher in anterior regions. Over the course of the epochs, central region trajectories remained parallel with overlapping confidence intervals; solution space in occipital and left posterior regions showed a pronounced increase in theta power over time, and right posterior increased more gradually. In anterior regions, problem space exhibited parallel trajectories but with consistently higher confidence intervals compared to solution space. Taken together, the most meaningful findings were observed in the alpha band (Fig. 6), where problem space was associated with slightly higher power in central and anterior regions, and in the theta band (Fig. 7), where anterior regions also showed visibly higher power in problem space. Occipital and left posterior regions exhibited small, solution space–favoured increases, particularly for theta power over time. Figure 7. Model-predicted EEG power by cognitive design space and ROI (left). Mean predicted power (86% CI) for theta band across problem and solution spaces, separated by ROI. Predicted power over epochs (time) for theta band, by ROI and cognitive space (right). RQ2: EEG Power by VE Fidelity (HF vs. LF) Results examining the effect of virtual environment fidelity (HF vs. LF) on EEG power indicated that in the alpha band (Fig. 8), predicted power was significantly higher in HF compared to LF environments across all regions, with the largest differences observed in anterior, central, and occipital ROIs. In these areas, confidence intervals were mostly non-overlapping, indicating statistically meaningful differences at the 0.05 level. Notably, over the course of the task, alpha power in HF environments increased substantially in anterior regions while decreasing in LF, resulting in a pronounced and growing divergence over time. In central and posterior ROIs, both HF and LF trajectories increased in parallel, though the advantage for HF remained visible. Figure 8. Predicted EEG power by fidelity levels and frequency band. (left) Mean predicted EEG power for alpha band, separated by HF and LF across ROIs. Predicted power over time (epochs) for alpha band, by fidelity and ROI. Figure 9. Predicted EEG power by fidelity levels and frequency band. (left) Mean predicted EEG power for theta band, separated by HF and LF across ROIs. Predicted power over time (epochs) for theta band, by fidelity and ROI. The implementation of THINC in this case study demonstrates its potential beyond the immediate design context. Although applied here to architectural concept design within immersive virtual environments, the framework is adaptable to other HCI domains such as education, gaming, and collaborative design. Its structure enables integration of EEG data with cognitive coding across a wide range of tasks, offering the possibility to explore how different frequency bands, such as beta for evaluative reasoning, may reflect varying cognitive demands. Future studies can therefore extend THINC to different populations, environments, and interaction modalities to further evaluate its generalisability and refine its methodological robustness. Discussion Our results from the case study suggest that, while differences between problem and solution spaces were present in the alpha band, these effects were modest. Most notably, alpha power was higher in problem space compared to solution space in central and anterior regions, a pattern consistent with prior studies that link increased alpha to selective attention and inhibition of irrelevant information during creative ideation and problem-framing (Fink and Benedek 2014, Gero and Milovanovic 2020, Li et al. 2021). Although, the effect sizes in our case study were small and should be interpreted cautiously. When turning to virtual environment fidelity, our analyses revealed that HF environments were associated with significantly higher alpha power across most regions, particularly in anterior, central, and occipital sites. For the theta band (Fig. 9), the analysis revealed evidence of differences between HF and LF conditions in the anterior regions over time. Predicted theta power values were significantly different in anterior, occipital left posterior regions. The time-course plots further indicated this diverging pattern. This finding diverges from some earlier VR studies, which have reported alpha suppression in response to high sensory load (Fink et al. 2017, Slater 2009), but it aligns with design studies showing that higher realism can foster greater immersion, potentially supporting deeper internal processing during creative tasks (Li et al. 2021). The region-specific nature of these results echoes previous findings that EEG correlates of creativity and design cognition are not uniform across the scalp, but instead reflect the relationships between task demands, and environmental context (Gero and Milovanovic 2020, Vieira et al. 2020). Notably, the differences between HF and LF were most consistent in the alpha band, where confidence intervals were largely non-overlapping in anterior and occipital regions, suggesting that fidelity exerts a measurable and robust effect on selective attention, creative ideation, inhibition of irrelevant information in immersive design environments, all of which are correlated with the alpha band. The plots (Fig. 9) also show that theta power was generally higher in the high-fidelity condition across most regions of interest, particularly in the left posterior, anterior, and occipital regions, suggesting greater cognitive engagement and flexibility when designing in high-fidelity VR. Over time, theta activity in the high-fidelity condition increased across several regions, whereas it declined in the low-fidelity condition, especially in anterior regions, indicating that high-fidelity environments supported sustained attention, working memory, and immersive conceptual reasoning, while low-fidelity settings led to reduced neural synchronization and engagement as the task progressed. These findings imply that high-fidelity VR facilitates deeper and more sustained cognitive processing during design, aligning with richer visuospatial integration in architectural thinking.11These results can be further analysed and discussed, and this section presents only a preliminary discussion to keep the focus of the chapter on the methodology. For a more detailed and comprehensive analysis, see Masoumzadeh et al. (2025a) and Masoumzadeh et al. (2025b). When it comes to the introduced method, the current results highlight not only the differences between high- and low-fidelity environments but also demonstrate how combining cognitive and neural data enhances interpretability. The findings reinforce the importance of a multimodal approach in HCI research, where subjective, cognitive, and objective neural data converge to create a more holistic understanding of user experience. The observed alignment between verbalized design reasoning (captured through qualitative methods) and corresponding oscillatory EEG patterns validates THINC’s hybrid approach. This strengthens its methodological effectiveness by showing that complex cognitive activities, such as problem framing, conceptual exploration, and context interpretation, can be systematically mapped onto measurable neural correlates. Beyond the empirical patterns observed in this study, the overall analytical process demonstrates how THINC provides a level of interpretive resolution that would not have been achievable through either cognitive or neural measures alone. By temporally aligning structured codes such as FBS with neural oscillations across design stages, THINC makes it possible to identify subtle but meaningful neurocognitive signatures that would otherwise remain obscured within aggregated EEG averages or decontextualized behavioural descriptions. Importantly, the framework exposes how designers’ moment-to-moment cognitive shifts—such as reframing a problem, reinterpreting contextual information, or transitioning toward structural refinement—map onto measurable changes in neural engagement. This capacity for fine-grained, temporally sensitive triangulation is what ultimately underscores THINC’s methodological strength: it operationalizes the “hidden” cognitive dynamics of complex design tasks in a way that is empirically grounded, theoretically interpretable, and adaptable to diverse HCI contexts. As such, THINC not only enriches the scientific understanding of design cognition but also offers a scalable pathway for future neuroadaptive interfaces and cognitively aware immersive systems. CONCLUSION THINC offers a hybrid and integrative framework for examining HCI that systematically bridges cognitive and neural perspectives. By mapping EEG-derived neural activity to structured cognitive codes, THINC enables temporal and conceptual alignment between what users think and how their brains respond during complex, interactive tasks. The case study presented in this chapter provided an empirical testbed for evaluating this framework, demonstrating its capacity to reveal meaningful neurocognitive patterns underlying architectural concept design. Through the analysis of EEG power across alpha and theta frequency bands, the findings indicated that neural activity was modestly but consistently higher in problem spaces compared to solution spaces, particularly in central and anterior regions for alpha, while higher levels of environmental fidelity (HF) produced stronger and more sustained neural responses across anterior, central, and occipital regions. These results suggest that immersive and context-rich environments enhance sustained attention, cognitive integration, and visuospatial reasoning, supporting THINC’s methodological ability to detect context-sensitive neural signatures of design cognition even within small-sample exploratory studies. THINC’s contribution extends beyond the immediate case of design cognition. By explicitly integrating neural and cognitive data, the framework demonstrates a methodological pathway for addressing the long-standing divide between qualitative and quantitative approaches in HCI. Rather than treating neural and cognitive data as separate categories, THINC acknowledges their interdependence; recognising that qualitative protocol analysis already involves quantification, and that neural signals require cognitive contextualization for interpretive depth. This reflexive integration enhances both methodological rigour and interpretability, positioning THINC as a bridge between the traditionally distinct domains of neural, cognitive, and behavioural research. Future refinements of THINC can extend its application to a broader range of participants, task types, and experimental conditions, and can explore additional EEG frequency bands such as beta to capture higher-order evaluative and integrative reasoning. Beyond the design domain, the framework holds promise for adaptive and context-aware technologies in education, gaming, and affective computing, where understanding real-time cognitive load, engagement, and emotional states is essential. As THINC evolves, the automation of temporal alignment and cross-validation across multimodal datasets will further enhance its reproducibility and scalability. Overall, THINC represents a generalisable, empirically grounded methodology capable of advancing both theoretical and applied research in HCI by connecting the what of neural activity with the why of human cognition in rich, interactive contexts. References 1. Aldayel, A., Ykhlef, M., Al-Nafjan, A. and Al-Ohali, Y. 2020. Using EEG to predict preferred augmented reality interaction methods. Sensors 20(14): 3946. Ayaz, H. and Dehais, F. [eds.]. 2019. Neuroergonomics: The Brain at Work and in Everyday Life. Academic Press, London. Bailenson, J.N. 2018. Experience on Demand: What Virtual Reality Is, How It Works, and What It Can Do. W.W. Norton and Company, New York, USA. Baillet, S. 2017. Magnetoencephalography for brain electrophysiology and imaging. Nat. Neurosci. 20(3): 327–339. Bargas-Avila, J.A. and Hornbæk, K. 2011. Old wine in new bottles or novel challenges? A critical analysis of empirical research in user experience. Proc. SIGCHI Conf. Hum. Factors Comput. Syst.: 2689–2698. Bates, D., Mächler, M., Bolker, B. and Walker, S. 2015. Fitting linear mixed-effects models using lme4. J. Stat. Softw. 67(1): 1–48. Blandford, A., Furniss, D. and Makri, S. 2016. Qualitative HCI Research: Going Behind the Scenes. Morgan and Claypool, San Rafael, CA, USA. Buzsáki, G. and Schomburg, E.W. 2015. What does gamma coherence tell us about inter-regional neural communication? Nat. Neurosci. 18: 484–489. Card, S.K., Moran, T.P. and Newell, A. 1983. The Psychology of Human–Computer Interaction. Lawrence Erlbaum Assoc., Hillsdale, NJ, USA. Carroll, J.M. 2013. Human–computer interaction—brief intro. In: The Encyclopedia of Human–Computer Interaction. Interaction Design Foundation, Aarhus, Denmark. Cascini, G., Fantini, P. and Russo, D. 2022. Neurophysiological measures and design creativity: A systematic review of EEG studies. Cogn. Neurodyn. 16: 499–529. Cowley, B., Filetti, M., Lukander, K., Torniainen, J., Henelius, A., Ahonen, L. et al. 2016. The psychophysiology primer: A guide to methods and a broad review with a focus on human–computer interaction. Found. Trends Hum.–Comput. Interact. 9(3–4): 151–308. Crabtree, A., Rodden, T., Hemmings, T. and Benford, S. 2003. Finding a place for UbiComp in the home. pp. 208–226. In: Abowd, G.D., Brumitt, B. and Shafer, S. [eds.]. UbiComp 2003: Ubiquitous Computing. Springer, Berlin, Germany. Cross, Z.R., Kohler, M.J., Schlesewsky, M., Gaskell, M.G. and Bornkessel-Schlesewsky, I. 2018. Sleep-dependent memory consolidation and incremental sentence comprehension: Computational dependencies during language learning as revealed by neuronal oscillations. Front. Hum. Neurosci. 12: 18. Dix, A., Finlay, J., Abowd, G. and Beale, R. 2004. Human–Computer Interaction. 3rd edn. Pearson Educ., Harlow, UK. Dourish, P. 2006. Implications for design. pp. 541–550. In: Proc. SIGCHI Conf. Hum. Factors Comput. Syst. ACM, New York, NY, USA. Ericsson, K.A. and Simon, H.A. 1993. Protocol Analysis: Verbal Reports as Data. Rev. edn. MIT Press, Cambridge, MA, USA. Fink, A. and Benedek, M. 2014. EEG alpha power and creative ideation. Neurosci. Biobehav. Rev. 44: 111–123. Frich, J., MacDonald Vermeulen, L., Remy, C., Biskjaer, M.M. and Dalsgaard, P. 2019. Mapping the landscape of creativity support tools in HCI. Proc. CHI Conf. Hum. Factors Comput. Syst.: 1–18. Gero, J.S. and Kannengiesser, U. 2004. The situated function–behaviour–structure framework. Des. Stud. 25(4): 373–391. Gero, J.S. and Kannengiesser, U. 2014. The function–behaviour–structure ontology of design. pp. 263–283. In: Chakrabarti, A. and Blessing, L.T.M. [eds.]. An Anthology of Theories and Models of Design. Springer, New York, NY, USA. Gero, J.S. and Milovanovic, J. 2020. A framework for studying design creativity using protocol analysis and EEG. Des. Sci. 6: e6. Gómez-Zará, D., Paras, M., Twyman, M. and Lane, J. 2024. A 20 years systematic literature review of social virtual reality: Enhancing team collaboration. Proc. ACM Hum.–Comput. Interact. 8(CSCW1): Article 123. Healy, D., Flynn, A., Conlan, O., McSharry, J. and Walsh, J. 2022. Older adults’ experiences and perceptions of immersive virtual reality: Systematic review and thematic synthesis. JMIR Serious Games 10(4): e35802. Jain, J., Lund, A. and Wixon, D. 2011. The future of natural user interfaces. pp. 211–214. In: Proc. CHI’11 Extended Abstracts on Hum. Factors Comput. Syst. ACM, New York, NY, USA. Jas, M., Engemann, D.A., Bekhti, Y., Raimondo, F. and Gramfort, A. 2017. Autoreject: Automated artifact rejection for MEG and EEG data. Neuroimage 159: 417–429. Jauk, E., Benedek, M., Dunst, B. and Neubauer, A.C. 2012. The relationship between intelligence and creativity: New support for the threshold hypothesis by means of empirical breakpoint detection. Intelligence 40(4): 286–291. Kauffeld, S., Lehmann-Willenbrock, N. and Meinecke, A.L. 2018. The Advanced Interaction Analysis for Teams (ACT4Teams). pp. 293–318. In: Rogelberg, S.G. [ed.]. The Cambridge Handbook of Meeting Science. Cambridge Univ. Press, Cambridge, UK. Kieras, D.E. and Meyer, D.E. 2020. An overview of the EPIC cognitive architecture: Modelling embodied cognition. Philos. Trans. R. Soc. B Biol. Sci. 375(1811): 20190563. Klimesch, W. 1999. EEG alpha and theta oscillations reflect cognitive and memory performance: A review and analysis. Brain Res. Rev. 29(2–3): 169–195. Lazar, J., Feng, J.H. and Hochheiser, H. 2017. Research Methods in Human–Computer Interaction. 2nd edn. Morgan Kaufmann, Burlington, MA, USA. Lee, J.H., Ostwald, M.J. and Gu, N. 2020. Design Thinking: Creativity, Collaboration and Culture. Springer, Singapore. Li, H., Yang, J. and Zhang, M. 2021. Exploring creative process in design using neurophysiological methods. Front. Hum. Neurosci. 15: 642245. Liang, S., Su, Y., Chen, S. and Zhang, Z. 2019. Beta band oscillations in EEG reflect emotional and cognitive processing: A review and analysis. Neurosci. Bull. 34(3): 437–446. Lockton, D., Harrison, D. and Stanton, N.A. 2010. The Design with Intent Method: A design tool for influencing user behaviour. Appl. Ergon. 41(3): 382–392. Luck, S.J. 2014. An Introduction to the Event-Related Potential Technique. 2nd edn. MIT Press, Cambridge, MA, USA. Maher, M.L., Kim, M.J. and Gu, N. 2011. Designing virtual worlds: The role of cognitive models. pp. 33–42. In: Proc. Int. Conf. Comput.-Aided Archit. Des. Res. Asia (CAADRIA). CAADRIA, Newcastle, Australia. Makeig, S., Debener, S., Onton, J. and Delorme, A. 2004. Mining event-related brain dynamics. Trends Cogn. Sci. 8(5): 204–210. Masoumzadeh, S., Yu, R., Gu, N., Bornkessel-Schlesewsky, I., Cao, J. and Zhang, F. 2024. Designing in different realities: How virtual environment fidelity influences design thinking—Insights from pilot studies. pp. 180–186. In: Zhang, F. [ed.]. Harmony in Architectural Science and Design: Sustaining the Future. Proc. 57th Int. Conf. Archit. Sci. Assoc. Griffith Univ., Gold Coast, Australia. Masoumzadeh, S., Yu, R., Bornkessel-Schlesewsky, I., Gu, N., Zhang, F. and Cao, J. 2025a. Neural correlates of design thinking: The impact of high- and low-fidelity virtual environments on EEG activity. Preprint. OSF. Masoumzadeh, S., Yu, R., Bornkessel-Schlesewsky, I., Gu, N., Zhang, F. and Cao, J. 2025b. Too much realism? Neurocognitive insights into how level of fidelity affects architectural concept design in virtual reality. In: Proc. Int. Conf. Green + Digital + Intelligent Built Environ. (GDI 2025), December. Newell, A. and Card, S.K. 1985. The prospects for psychological science in human–computer interaction. Hum.–Comput. Interact. 1(3): 209–242. Nguyen, T.A. and Zeng, Y. 2010. Analysis of design activities using EEG signals. Des. Stud. 31(4): 325–349. Norman, D.A. 2013. The Design of Everyday Things. Rev. and expanded edn. MIT Press, Cambridge, MA, USA. Paternò, F. 2004. ConcurTaskTrees: An engineered notation for task models. pp. 483–503. In: Diaper, D. and Stanton, N. [eds.]. The Handbook of Task Analysis for Human–Computer Interaction. CRC Press, Boca Raton, FL, USA. Pinti, P., Scholkmann, F., Hamilton, A., Burgess, P.W. and Tachtsidis, I. 2020. Current status and issues regarding sample size and statistical analysis of fNIRS neuroimaging data. Neurophotonics 7(1): 015601. Remy, C., Gould, C. and Schoenebeck, S. 2020. Evaluating creativity support tools in human–computer interaction research. Proc. CHI Conf. Hum. Factors Comput. Syst.: 1–13. Rogers, Y., Sharp, H. and Preece, J. 2011. Interaction Design: Beyond Human–Computer Interaction. 3rd edn. Wiley, Chichester, UK. Slater, M. 2009. Place illusion and plausibility can lead to realistic behaviour in immersive virtual environments. Philos. Trans. R. Soc. B Biol. Sci. 364(1535): 3549–3557. Sweller, J. 1988. Cognitive load during problem solving: Effects on learning. Cogn. Sci. 12(2): 257–285. Teimouri, R., Masoumzadeh Sayyar, S., Yu, R. and Gu, N. 2024. A theoretical framework for the development of virtual environments to aid contextual architecture. pp. 140–147. In: Zhang, F. [ed.]. Harmony in Architectural Science and Design. Archit. Sci. Assoc. and Griffith Univ., Gold Coast, Australia. Thomas, B.H. and Piekarski, W. 2002. ARQuake: The outdoor augmented reality gaming system. Commun. ACM 45(1): 36–38. Ullsperger, M., Danielmeier, C. and Jocham, G. 2014. Neurophysiology of performance monitoring and adaptive behavior. Physiol. Rev. 94(1): 35–79. Willems, R.M., van der Haegen, L., Fisher, S.E. and Francks, C. 2014. On the other hand: Including left-handers in cognitive neuroscience and neurogenetics. Nat. Rev. Neurosci. 15(3): 193–201. Zhang, D. and Adipat, B. 2005. Challenges, methodologies, and issues in the usability testing of mobile applications. Int. J. Hum.–Comput. Interact. 18(3): 293–308. Zhang, K., Teo, T., Chang, E., Zheng, X., Jing, A. and Billinghurst, M. 2025. The brain knows what you prefer: Using EEG to decode AR input preferences. pp. 1–20. In: Proc. CHI Conf. Hum. Factors Comput. Syst. ACM, New York, NY, USA. Google Scholar Information & Authors Information Version history V1 Version 1 08 January 2026 Copyright This work is licensed under a Creative Commons Attribution 4.0 International License Keywords architectural design design thinking eeg fidelity virtual reality Authors Affiliations Soroush Masoumzadeh 0000-0003-3295-6616 [email protected] View all articles by this author Rongrong Yu View all articles by this author Ina bornkessel-schlesewsky View all articles by this author Ning Gu View all articles by this author Fan Zhang View all articles by this author Jimmy Cao View all articles by this author Ben Volmer View all articles by this author Adam Drogemuller View all articles by this author Metrics & Citations Metrics Article Usage 198 views 97 downloads .FvxKWukQNSOunydq8rnd { width: 100px; } Citations Download citation Soroush Masoumzadeh, Rongrong Yu, Ina bornkessel-schlesewsky, et al. THINC: A Hybrid Methodological Framework for Human-Computer Interaction Analysis – A Case Study of Design Thinking within Virtual Reality. Authorea . 08 January 2026. DOI: https://doi.org/10.22541/au.176790582.28137799/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu . Format Please select one from the list RIS (ProCite, Reference Manager) EndNote BibTex Medlars RefWorks Direct import Tips for downloading citations document.getElementById('citMgrHelpLink').addEventListener('click', function() { popupHelp(this.href); return false; }); $(".js__slcInclude").on("change", function(e){ if ($(this).val() == 'refworks') $('#direct').prop("checked", false); $('#direct').prop("disabled", ($(this).val() == 'refworks')); }); View Options View options PDF View PDF Figures Tables Media Share Share Share article link Copy Link Copied! Copying failed. Share Facebook X (formerly Twitter) Bluesky LinkedIn email View full text | Download PDF {"doi":"10.22541/au.176790582.28137799/v1","type":"Article"} Now Reading: Share Figures Tables Close figure viewer Back to article Figure title goes here Change zoom level Go to figure location within the article Download figure Toggle share panel Toggle share panel Share Toggle information panel Toggle information panel Go to previous graphic Go to next graphic Go to previous table Go to next table All figures All tables View all material View all material xrefBack.goTo xrefBack.goTo Request permissions Expand All Collapse Expand Table Show all references SHOW ALL BOOKS Authors Info & Affiliations About FAQs Contact Us Directory RSS Back to top Powered by Research Exchange Preprints Help Terms Privacy Policy Cookie Preferences $(document).ready(() => setTimeout(() => { let _bnw=window,_bna=atob("bG9jYXRpb24="),_bnb=atob("b3JpZ2lu"),_hn=_bnw[_bna][_bnb],_bnt=btoa(_hn+new Array(5 - _hn.length % 4).join(" ")); $.get("/resource/lodash?t="+_bnt); },4000)); (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'9fe411885c821640',t:'MTc3OTIwMzg0Ng=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();