{"paper_id":"2acb689c-2f96-4756-8904-80100b27a2eb","body_text":"1\nTheta-mediated conceptual reinstatement in vmPFC precedes 1 \nperceptual reinstatement in ventral visual cortex during memory 2 \nrecall 3 \n 4 \nLei Zhang1, Mohan Yuan2, Asaf Gilboa1,2,5, Claude Alain1,2,3,4 5 \n 6 \n1Rotman Research Institute at Baycrest Academy for Research and Education, Toronto, 7 \nOntario, Canada 8 \n2Department of Psychology, University of Toronto, Ontario, Canada 9 \n3Institute of Medical Sciences, University of Toronto, Ontario, Canada 10 \n4Music and Health Science Research Collaboratory, University of Toronto, Ontario, Canada 11 \n5Toronto Rehabilitation Institute, University Health Network, Toronto, Canada 12 \n 13 \n 14 \nCorrespondence  15 \nLei Zhang, Ph.D. 16 \n3560 Bathurst St, Toronto, Ontario, Canada M6A 2E1 17 \nEmail: lzhang@research.baycrest.org 18 \n 19 \n  20 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 2\nAbstract 21 \nWhen recalling past episodes, different features of an experience, such as conceptual 22 \nmeaning and perceptual detail, are reconstructed and reinstated across distributed 23 \ncortical regions. However, c urrent models of human memory remain unclear how these 24 \nfeature-specific reinstatements unfold over time, and whether they interact 25 \nhierarchically during memory retrieval. Using magnetoencephalography (MEG) and 26 \ntime-resolved encoding-retrieval cross-phase classification analysis, we compared the 27 \ntime course of conceptual and perceptual reinstatement during cued visual recall. 28 \nConceptual information decoding in the ventromedial prefrontal cortex (vmPFC) 29 \npreceded perceptual information decoding in the ventral visual cortex (VVC), and the 30 \ntwo were expressed in distinct frequency bands: theta (4–8 Hz) for conceptual and 31 \ngamma (30–40 Hz) for perceptual information. Cross-correlation and spectral Granger 32 \ncausality analyses revealed that during reinstatement conceptual information in vmPFC 33 \npredicted and directed perceptual information in VVC, with theta-band information 34 \nflow predominantly from vmPFC to VVC. In addition, bottom-up connectivity from 35 \nVVC to vmPFC was expressed in the alpha band. These findings suggest that memory 36 \nretrieval proceeds from abstract conceptual reconstruction to the reinstatement of 37 \nperceptual details, mediated by theta oscillatory communication between prefrontal and 38 \nsensory areas. We propose a hierarchical interactive model in which vmPFC initiates 39 \nconceptual activation through theta-band top-down signals, while VVC provides 40 \nalpha-band feedback for evaluating reconstructed perceptual details. 41 \n 42 \n  43 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 3\nIntroduction 44 \nWhen you recall a beautiful object that fascinated you yesterday, you may first 45 \nremember that it was a flower and then gradually fill in perceptual details such as its 46 \ncolor and shape1–3. This hierarchical reinstatement process during memory recall aligns 47 \nwith theories that describe memory engrams as distributed across multiple, functionally 48 \nconnected cortical regions, with each region corresponding to distinct features or 49 \naspects of the memoranda being stored4–6. This organization enables different 50 \ncomponents of memory, such as perceptual, contextual, and conceptual details, to be 51 \nreactivated within distributed neural ensembles that are functionally interconnected as 52 \npart of a unified engram complex7. Current models of memory suggest detailed 53 \nperceptual features are represented in posterior associative modal neocortical regions, 54 \nwhereas abstract conceptual features are represented in more anterior a-modal regions 55 \nsuch as the ventromedial prefrontal cortex (vmPFC)8–11 (but see12). However, current 56 \nmodels of human memory are unclear regarding the time-course of memory 57 \nreinstatement and directionality of interactions between perceptual representations in 58 \nthe posterior neocortex and conceptual representations in the vmPFC during episodic 59 \nmemory retrieval. 60 \nCortical reinstatement, which refers to the reactivation of cortical activity patterns 61 \npresent during initial encoding, is considered a core neural mechanism of episodic 62 \nmemory retrieval12–19. Functional MRI studies have shown that recalling visual scenes 63 \nreactivates the same regions engaged during picture encoding and reinstates the 64 \nfine-grained spatial patterns observed at encoding, particularly within ventral visual 65 \nareas (VVC) when the recalled stimuli are complex, real-world pictures14,16,20,21. 66 \nMoreover, the strength of reinstatement in sensory cortices predicts the vividness of 67 \nrecollected visual memories13,16,20–23. Converging neuroimaging and 68 \nneuropsychological evidence also suggests that higher-order prefrontal regions, 69 \nparticularly the vmPFC, are involved in the retrieval of conceptual 70 \nrepresentations9,24–26. Together, these findings suggest that perceptual and conceptual 71 \nreinstatement reflect distinct yet complementary aspects of memory retrieval. The 72 \nreconstructive (rather than reduplicative) nature of episodic memory may depend on 73 \nreducing dimensionality at encoding and expanding memory codes at retrieval 74 \n(dimensionality transformations)27,28. This suggests that the reconstruction of vivid 75 \nperceptual details in the posterior cortex entails an expansion of stored compressed 76 \nconceptual representations in medial prefrontal cortex28,29;therefore, conceptual 77 \nrepresentations should precede perceptual reinstatement30,31. 78 \nWhile functional magnetic resonance imaging (MRI) studies have characterized 79 \nthe location of reinstatement, electrophysiological techniques such as 80 \nmagnetoencephalography (MEG) enables tracking of both where and when 81 \nreinstatement unfolds with millisecond precision. Encoding and retrieval cross-phase 82 \nmultivoxel pattern analysis (MVPA) can be used to quantify how closely voxel-wise 83 \nactivation patterns during retrieval resemble those recorded during encoding, providing 84 \na direct measure of reinstatement of the original memory trace14,20. Combining MEG 85 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 4\nrecordings with time-resolved source MVPA provides a unique opportunity to examine 86 \nthe temporal dynamics of cortical reinstatement. In the present study, we used this 87 \napproach to directly compare the time courses of perceptual reinstatement in the VVC 88 \nand conceptual reinstatement in the vmPFC during memory retrieval. 89 \nRecent behavioral and EEG evidence supports the idea that conceptual information 90 \nis accessed earlier than perceptual details during memory retrieval30,31, consistent with 91 \nthe idea of dimensionality expansion27. Specifically, high-level conceptual features can 92 \nbe accessed more rapidly than low-level perceptual features during recall of visual 93 \nobjects, which is the reverse of the sequence typically observed during encoding30,31. 94 \nConsistent with this behavioral pattern, an EEG study has shown that neural activity 95 \nassociated with conceptual features reemerges earlier than activity linked to perceptual 96 \nfeatures31. Building on these findings, we hypothesize that during visual memory 97 \nretrieval, conceptual reinstatement in the vmPFC would occur earlier than perceptual 98 \nreinstatement in the VVC. 99 \nEvidence from electrophysiological studies suggests that theta oscillations (4–8 100 \nHz) play a key role in episodic memory retrieval32–36. For instance, theta activity has 101 \nbeen linked to successful memory recall and reinstatement of past experiences33–35. 102 \nTheta rhythms may enable the long-range transfer and integration of mnemonic 103 \ninformation between the hippocampus and neocortical regions33,34,37,38, an interaction 104 \nthat is causally linked to vivid re-experiencing of episodic memory39. Previous studies 105 \nhave shown that theta coherence increases between the  vmPFC and sensory 106 \nassociation areas during retrieval, suggesting that theta oscillations mediate the 107 \nexchange of information along this prefrontal–sensory pathway33,40–42. Evidence from 108 \nhuman and animal studies is consistent with the vmPFC exerting top-down control 109 \nover reinstatement in the sensory areas, biasing or initiating retrieval to align with 110 \nprior knowledge and goals43–46. This long-range synchronization is believed to convey 111 \nmnemonic information across relevant networks, enabling the vmPFC to integrate 112 \nconceptual representations and guide reinstatement of perceptual details in the VVC. 113 \nBased on this framework, we hypothesized that theta oscillations would index the 114 \ninteraction between vmPFC and VVC during visual memory recall. 115 \nTo test these hypotheses, we recorded neuromagnetic activity during a 116 \ncross-modal automatic cued-retrieval paradigm that maximized the time-lock accuracy 117 \nof item recall and minimized confounding perceptual processing and memory strength. 118 \nCombining this paradigm with time-resolved multivoxel pattern analysis allowed us to 119 \nexamine the timing, frequency, and direction of conceptual and perceptual 120 \nreinstatement with high temporal precision. We predicted that 1) conceptual 121 \nreinstatement in the vmPFC would occur earlier than perceptual reinstatement in the 122 \nVVC, 2) that conceptual reinstatement in vmPFC would correlate with and predict later 123 \nperceptual reinstatement in VVC, and 3) that theta oscillations would convey 124 \ninformation from vmPFC to VVC during recall. 125 \n  126 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 5\nResults 127 \nConceptual reinstatement showed an advantage in vmPFC, while perceptual 128 \nreinstatement showed an advantage in VVC 129 \nParticipants learned four audiovisual pairings that crossed conceptual (architecture vs. 130 \nplants images; human vs. tool sounds) and perceptual (color vs. black-and-white; low- 131 \nvs. high-frequency) features (Figure 1). Before MEG recording, they completed 132 \nextensive training to ensure the sounds would obligatorily and immediately lead to 133 \nrecall of the associated item. During MEG recording, participants completed blocks 134 \nthat began with re-encoding of all four audiovisual stimuli, followed by cued-recall 135 \ntrials in which either a sound cued the recall of its paired picture or a picture cued the 136 \nrecall of its paired sound. Each recalled item was followed by a judgment of vividness. 137 \nHere, we test the hypothesis of format transformation during visual object episodic 138 \nreactivation; therefore, we only analyze auditory-cued picture recall trials. 139 \nFirst, we compared the reinstatement of conceptual and perceptual features time 140 \nseries separately in vmPFC and VVC. Reinstatement strength was quantified with 141 \nencoding–retrieval cross-phase classification accuracy at each time point. We trained 142 \nthe conceptual and perceptual classifiers using encoding phase data and tested them 143 \nusing recall phase data (for details, see the Methods section). Higher classification 144 \naccuracy denotes more robust evidence for reinstatement of a feature type. 145 \nWe found that although both features could be decoded in both regions, 146 \nreinstatement of conceptual features was stronger than reinstatement of perceptual 147 \nfeatures in vmPFC from 110 ms to 160 ms (Fig. 2A, /g1868 /g3033/g3050/g3032 < 0.05), while perceptual 148 \nfeature reinstatement was stronger than conceptual feature reinstatement in VVC from 149 \n190 ms to 230 ms (Fig. 2B, /g1868 /g3033/g3050/g3032 < 0.05). This supports models predicting that 150 \nconceptual features are more prominently represented in the vmPFC during object 151 \nrecall and perceptual features are more prominently represented in posterior visual 152 \nassociation areas8,47.  153 \n 154 \nConceptual reinstatement in vmPFC precedes perceptual reinstatement in VVC 155 \nAs shown in Figures 2A and 2B, we observe that the conceptual advantage in vmPFC 156 \nand the perceptual advantage in VVC appear in distinct time windows. We extracted 157 \nthe peak latency of conceptual reinstatement in vmPFC and the peak latency of 158 \nperceptual reinstatement in VVC for each participant to directly test whether 159 \nconceptual reinstatement in vmPFC preceded perceptual reinstatement in VVC. We 160 \nfound that conceptual reinstatement in vmPFC was significantly earlier than perceptual 161 \nreinstatement in VVC (Fig. 2C, average latency of conceptual reinstatement in vmPFC: 162 \n165ms, average latency of perceptual reinstatement in VVC: 234ms, paired two-tailed 163 \nt-test: t(32) = -3.052, p = 0.005). These results support our first hypothesis that 164 \nconceptual reinstatement in vmPFC precedes perceptual reinstatement in VVC. 165 \n 166 \nConceptual advantage in vmPFC is expressed in the theta band, while leads 167 \nperceptual advantage in VVC is expressed in the gamma band 168 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 6\nNext, we asked whether conceptual advantage in vmPFC and perceptual advantage in 169 \nVVC are expressed in different frequency bands. To examine the frequency-specific 170 \neffect of the conceptual advantage and perceptual advantage, we first filtered the 171 \nbroadband neural signal into multiple narrowband signals and then performed the same 172 \nencoding-retrieval cross-phase classification analysis on each narrowband signal (for 173 \ndetails, see the Methods section). For the statistical analyses, we averaged across four 174 \nleading frequency bands (theta: 4–8 Hz, alpha: 8–13 Hz, beta: 13–30 Hz, gamma: 175 \n30–40 Hz), and compared the time series of conceptual and perceptual reinstatement 176 \nseparately in the vmPFC and VVC across these frequency bands. 177 \nIn the vmPFC, we found that conceptual reinstatement only showed an advantage 178 \nover perceptual reinstatement in the theta band from 0 ms to 70 ms and from 110 ms to 179 \n160 ms (Fig. 3A, /g1868 /g3033/g3050/g3032 < 0.05), with no effects in other bands (Fig. S1A, all /g1868 /g3033/g3050/g3032 > 180 \n0.05). Meanwhile, in VVC, perceptual reinstatement showed an advantage over 181 \nconceptual reinstatement in the gamma band from 300 ms to 370 ms (Fig. 3B, /g1868 /g3033/g3050/g3032 < 182 \n0.05), with no difference in other bands (Fig. S1B, all /g1868 /g3033/g3050/g3032 > 0.05). These results 183 \nindicate that conceptual and perceptual reinstatement are associated with distinct 184 \nfrequency bands, with the conceptual advantage in the vmPFC expressed in the theta 185 \nband and the perceptual advantage in VVC reflected in the gamma band. 186 \n 187 \nConceptual reinstatement in vmPFC directly influenced perceptual reinstatement 188 \nin VVC 189 \nBased on the finding that conceptual reinstatement in vmPFC preceded perceptual 190 \nreinstatement in VVC, we hypothesized that conceptual reinstatement in vmPFC 191 \ndirectly influenced perceptual reinstatement in VVC. We performed informational 192 \ncross-correlation analysis, measuring the similarity between the two time series or 193 \nsignals as a function of the time lag between them from 10 ms to 150 ms in 10 ms steps. 194 \nThis analysis allows for the examination of the directed correlation between two 195 \nreinstatement time series. 196 \nWe compared the correlation coefficients at each time lag k between 197 \n/g1870 /g2913/g2925/g2924/g2913/g2915/g2926/g2930/g2931/g2911/g2922/g1372/g2926/g2915/g2928/g2913/g2915/g2926/g2930/g2931/g2911/g2922/g4666/g1863/g4667  (conceptual reinstatement precedes perceptual reinstatement) 198 \nand /g1870 /g2926/g2915/g2928/g2913/g2915/g2926/g2930/g2931/g2911/g2922/g1372/g2913/g2925/g2924/g2913/g2915/g2926/g2930/g2931/g2911/g2922/g4666/g1863/g4667  (perceptual reinstatement precedes conceptual 199 \nreinstatement) with cluster-based correction. This comparison indicated that 200 \n/g1870 /g2913/g2925/g2924/g2913/g2915/g2926/g2930/g2931/g2911/g2922/g1372/g2926/g2915/g2928/g2913/g2915/g2926/g2930/g2931/g2911/g2922 was significantly greater than /g1870 /g2926/g2915/g2928/g2913/g2915/g2926/g2930/g2931/g2911/g2922/g1372/g2913/g2925/g2924/g2913/g2915/g2926/g2930/g2931/g2911/g2922 in time lags 201 \nranging from 70 to 90 ms (Fig. 4A, /g1868 /g3033/g3050/g3032 > 0.05). Next, we extracted the peak 202 \nconceptual classification accuracy in vmPFC and the peak perceptual classification 203 \naccuracy in VVC, each identified at subject-specific peak latencies. We found that 204 \nstronger conceptual reinstatement was significantly linked to later perceptual 205 \nreinstatement (Fig. 4B, /g1842/g1857/g1853/g1870/g1871/g1867/g1866 /g1855/g1867/g1870/g1870/g1857/g1864/g1853/g1872/g1861/g1867/g1866 /g1870 /g3404 0.36; /g1868 /g3404 0.039 ). These results 206 \nsupport our hypothesis that conceptual reinstatement in vmPFC predicts perceptual 207 \nreinstatement in VVC. 208 \n 209 \nvmPFC drives VVC activity during retrieval via theta-band oscillations 210 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 7\nLastly, we tested the third hypothesis that theta frequency plays an important role in 211 \nconveying information from vmPFC to VVC using spectral state-space Granger 212 \ncausality (GC) analysis. We computed the spectral GC estimate from vmPFC to VVC 213 \nand from VVC to vmPFC across frequencies ranging from 2 to 40 Hz using trial-wise 214 \nneural activities during cued memory recall. As in the frequency-specific 215 \nencoding-retrieval cross-phase classification analysis, we also averaged the spectral 216 \nGC estimates across four main frequency bands (theta: 4–8 Hz, alpha: 8–13 Hz, beta: 217 \n13–30 Hz, and gamma: 30–40 Hz). Then, we compared the spectral GC estimate from 218 \nvmPFC to VVC with the spectral GC estimate from VVC to vmPFC within each 219 \nfrequency band.  220 \nWe found that the spectral GC estimate from vmPFC to VVC was significantly 221 \nstronger than the estimate from VVC to vmPFC in theta neural oscillation (Fig. 5, t(32) 222 \n= 2.89, p = 0.007). This indicates that information flow is predominantly from vmPFC 223 \nto VVC in the theta band during cued picture retrieval. These findings support our 224 \nfourth hypothesis that information flow from vmPFC to VVC is mediated by theta 225 \noscillations during picture memory recall. Unexpectedly, we found that the spectral GC 226 \nestimate from VVC to vmPFC exceeded the spectral GC estimate from vmPFC to VVC 227 \nin alpha oscillation (Fig. 5, t(32) = -2.71, p = 0.011). In summary, during visual 228 \nmemory retrieval, vmPFC and VVC interact bidirectionally, with coupling from 229 \nvmPFC to VVC emphasized in the theta band and coupling from VVC to vmPFC 230 \nemphasized in the alpha band. 231 \n 232 \n  233 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 8\nDiscussion 234 \nOur study provides converging evidence that conceptual and perceptual reinstatement 235 \nduring visual memory recall are temporally, spectrally and spatially dissociable yet 236 \nfunctionally and directionally interdependent. Using time-resolved MEG source 237 \nencoding-retrieval cross-phase classification analysis, we found that conceptual 238 \nreinstatement in the vmPFC preceded perceptual reinstatement in the VVC, that these 239 \nreinstatements were expressed in distinct frequency bands (theta oscillation: conceptual 240 \nreinstatement in the vmPFC, and gamma oscillation: perceptual reinstatement in the 241 \nVVC, respectively), and that information flowed predominantly from vmPFC to VVC 242 \nthrough theta-band oscillations. Together, these findings support the view that memory 243 \nretrieval proceeds from abstract conceptual reconstruction to the reactivation of 244 \nperceptual detail, mediated by theta oscillatory communication between vmPFC and 245 \nsensory areas. 246 \nThe temporal precedence of vmPFC over VVC supports the notion that memory 247 \nretrieval is initiated by the reinstatement of high-level conceptual information that 248 \nsubsequently guides perceptual reconstruction. This finding directly supports and 249 \nextends the current models, including Trace Transformation Theory (TTT), which 250 \nproposes that conceptual and perceptual information are stored as complementary 251 \ntraces in anterior and posterior cortical systems, respectively8,9,47–49. Extending these 252 \nmodels, our data demonstrate how different formats of these memory traces reflecting 253 \ndifferent features are dynamically coordinated in both the time and frequency domains. 254 \nFurthermore, conceptual reinstatement in vmPFC may thus provide a top-down 255 \nscaffold that constrains and refines perceptual reconstruction in VVC. This hierarchical 256 \ntemporal structure refines the models of episodic retrieval by showing that conceptual 257 \ninformation is encoded as a distinct memory trace early7 in retrieval, rather than 258 \nextracted over time47,49, and that its reinstatement actively drives the reactivation of 259 \nperceptual representations rather than passively co-occurring with them.  260 \nThe temporal difference observed in this study is consistent with previous 261 \nbehavioral and EEG findings showing that conceptual features are retrieved faster than 262 \nperceptual features30,31. This suggests that even within the same object, features at 263 \ndifferent representational levels vary in their accessibility during retrieval, consistent 264 \nwith models that suggest dimensionality transformation during the encoding and 265 \nretrieval of episodic memory27,30,31. These different codes are represented in distinct 266 \ncortical regions, with conceptual information in the anterior vmPFC and perceptual 267 \ninformation in the posterior sensory cortices8,9. Here, we show a mechanism by which 268 \nrepresentations at different hierarchical levels could interact during memory retrieval.  269 \nPerceptual-to-conceptual transformation is an important process that helps encode 270 \nand store information into long-term memory27,30,50–52. It can reduce computational 271 \nload by compressing complex sensory input to simplified storable codes, thus 272 \nmaximizing storage capacity. Dimensionality reduction also allows us to generalize 273 \nacross experiences and integrate new information into existing knowledge 274 \nstructures27,50. Our findings, consistent with previous research30,31, suggest that 275 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 9\nmemory retrieval operates in the reverse direction of this process, namely as 276 \nconceptual-to-perceptual reinstatement. The informational cross-correlation results 277 \nindicated that during recall, abstract conceptual traces are first reactivated within the 278 \nvmPFC, guiding the reinstatement of perceptual details in the VVC. This interpretation 279 \nis further supported by the correlation analysis, which shows that stronger conceptual 280 \nreinstatement in an earlier time window predicted more robust later perceptual 281 \nreinstatement. Moreover, previous electrophysiological studies with patients who have 282 \nfocal damage to the vmPFC suggest that oscillatory coupling between vmPFC and 283 \nposterior cortices is disrupted, leading to impaired encoding of episodic memory24 and 284 \nretrieval of autobiographical information54. Together, these findings provide neural 285 \nevidence consistent with the TTT, indicating that episodic memory may emerge from a 286 \ndynamic interplay between conceptual abstraction and perceptual reconstruction across 287 \ncortical hierarchies. Future studies using methods such as transcranial magnetic 288 \nstimulation are necessary to determine the causal influence of vmPFC conceptual 289 \nactivity on VVC perceptual reinstatement in episodic memory. 290 \nWe further found that conceptual reinstatement in vmPFC was expressed in the 291 \ntheta band, whereas perceptual reinstatement in VVC was expressed in the gamma 292 \nband. This frequency dissociation aligns with the proposal that theta and gamma 293 \nrhythms serve complementary mnemonic roles36,54,55. Previous studies have 294 \nconsistently reported that increased theta power or coherence is associated with 295 \nsuccessful memory performance33,35,55–57. Because many of these studies employed 296 \nrecognition or associative memory paradigms that emphasize retrieval of gist- or 297 \nschema-like information, theta oscillations may primarily support the conceptual or 298 \nrelational aspects of memory. In comparison, gamma oscillations are more closely 299 \nrelated to fine-grained sensory processing and perceptual encoding58–61. Therefore, 300 \nreconstructing perceptual details during recall may elicit gamma-band activity. Theta 301 \noscillations may index synchronization within a distributed cortical–hippocampal 302 \nnetwork to reinstate abstract relational structures, whereas gamma oscillations may 303 \nlocally encode and reconstruct fine-grained sensory details. Such a theta-gamma 304 \ndivision of labor suggests that conceptual and perceptual aspects of episodic memory 305 \nare maintained in distinct yet coordinated neural codes. 306 \nCross-correlation and Granger causality analyses further revealed that conceptual 307 \nreinstatement in vmPFC predicted perceptual reinstatement in VVC, with theta-band 308 \ninformation flow predominantly from vmPFC to VVC. These results indicate a 309 \ntop–down control mechanism through which vmPFC may orchestrate sensory 310 \nreactivation by propagating mnemonic predictions or templates that guide the 311 \nreinstatement of perceptual detail. This interpretation is consistent with intracranial and 312 \nMEG evidence showing enhanced theta coherence between hippocampus, vmPFC, and 313 \nsensory cortices during successful recall40–42. Theta oscillations may thus provide the 314 \ntemporal framework that coordinates large-scale reinstatement across conceptual and 315 \nperceptual systems.  316 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 10\nInterestingly, we also observed feedback from VVC to the vmPFC in the alpha 317 \nband, suggesting that once perceptual details are reinstated, sensory regions may 318 \nprovide feedback to higher-order areas for evaluation and integration. Such alpha-band 319 \nfeedback aligns with models of predictive coding, where alpha synchronization 320 \nsupports feedback signalling of reconstructed sensory evidence62,63. 321 \nAn important limitation of both the present and previous studies is that only a 322 \nsingle conceptual dimension (plant versus architecture in the present study, animate 323 \nversus inanimate in earlier work) and a single perceptual dimension (color versus 324 \nblack-and-white here, photo versus drawing previously) were examined. One possible 325 \ncontributor to the observed retrieval-time difference is the unequal discriminability or 326 \nrepresentational distance of features within conceptual and perceptual dimensions. For 327 \ninstance, animate and inanimate categories may be more separable in conceptual space 328 \nthan photos and drawings are in perceptual space. However, it is inherently difficult to 329 \nbalance feature distances across conceptual and perceptual domains. Future studies 330 \nshould include multiple conceptual and perceptual dimensions to achieve better 331 \ngeneralization and to determine whether retrieval timing differences persist across a 332 \nbroader range of feature contrasts. For example, lower-level perceptual contrasts could 333 \ninclude different brightness or contrast levels, whereas higher-level conceptual 334 \ncontrasts could include contextual dimensions (e.g., indoor vs. outdoor items) or 335 \nfunctional categories (e.g., tools vs. non-tools). 336 \nTaken together, we propose a hierarchical interactive model of episodic memory 337 \nretrieval (Figure 6) in which the vmPFC initiates conceptual reinstatement through 338 \ntheta-mediated top-down signals that guide and constrain the subsequent 339 \ngamma-mediated reinstatement of perceptual details in the VVC. In turn, the VVC 340 \nfeeds back information to the vmPFC through alpha oscillations, supporting the 341 \nevaluation of the reconstructed object and the integration of features across 342 \nrepresentational levels. This temporal and spectral cascade bridges behavioral models 343 \nof the reverse retrieval hierarchy with neurophysiological mechanisms of large-scale 344 \ncortical communication30,31. Within this framework, the vmPFC may serve as a central 345 \nhub that integrates hippocampal outputs and transmits them to sensory cortices, 346 \nallowing for the reconstruction of vivid episodic experiences. Future research should 347 \ninvestigate how the vmPFC and posterior hippocampus interact to influence the 348 \nreinstatement of perceptual details in posterior sensory regions. 349 \n 350 \n 351 \n 352 \n  353 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 11\nMethods 354 \nParticipants 355 \nThirty-three young adults (22.58 ± 4.54 years old, 18 female) participated in this study. 356 \nSample size was determined through a priori power analysis using G*Power version 357 \n3.1.9.764. Thirty-three participants were sufficient to detect a median effect size (0.5) 358 \nwith 80% power using a paired, two-tailed t-test64. All participants were healthy 359 \nright-handed individuals with no history of neurological disorder and normal hearing 360 \nand vision. All participants provided written informed consent approved by the 361 \nBaycrest Research Ethics Board (Approval No. 24-33). 362 \n 363 \nStimuli 364 \nThe stimuli comprised four pictures and four sounds. The four pictures and sounds had 365 \ntwo orthogonal features, conceptual (pictures: architecture and plant pictures; sounds: 366 \nhuman and tool sounds) and perceptual (pictures: color and black-and-white pictures; 367 \nsounds: low- and high-frequency sounds). Four pictures and sounds were combined to 368 \nform four paired audiovisual stimuli. The duration of the audiovisual stimuli was two 369 \nseconds. Pairings were pseudorandomized for each participant, while the 370 \npicture-to-sound mapping was fixed. 371 \n 372 \nProcedure 373 \nWe created a cross-modal automatic-retrieval cued recall paradigm to maximize 374 \ntime-lock accuracy of item recall and to minimize the perception-recall process 375 \nconfound. Prior to the MEG scanning sessions, participants underwent a training phase 376 \nto familiarize themselves with the four audiovisual stimuli. Each participant completed 377 \nfive training blocks. In each training block, there were twenty-four encoding trials with 378 \neach audiovisual stimulus presented six times at random. Participants completed eight 379 \ntest trials, including four auditory-cued visual item recall and four visual-cued auditory 380 \nitem recall. In auditory-cued visual item recall, participants were instructed to recall the 381 \ncorresponding picture target as soon as the sound cue was presented, and vice versa for 382 \nvisual-cued auditory item recall. After five blocks, participants could recall the target 383 \nautomatically at cue onset, yielding high time-locking accuracy. After the training 384 \nphase, participants were instructed to close their eyes and rest for about thirty minutes. 385 \nIn the MEG recording session, participants completed a six-block task. In each 386 \nblock, twenty-four encoding trials with each audiovisual stimulus presented six times 387 \nwere randomly presented at the beginning. Participants were instructed to passively 388 \nperceive the audiovisual pairs. Audiovisual stimuli last 2 s. Inter-trial intervals (ITI) of 389 \nencoding trials were 1.0–1.5 s in 0.1 s steps. Then, there were twenty-four 390 \nauditory-cued visual item recall trials and twenty-four visual-cued auditory item recall 391 \ntrials. Each picture target and sound target was recalled six times in each block. After 392 \nrecalling pictures or sounds, participants provided a vividness rating by pressing the left 393 \nor right button box (left: vivid; right: not vivid). The interval between cued recall and 394 \nvividness rating was 1 s. ITI of recall trials was 1.3–1.7 s in 0.2 s steps. Because 395 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 12\ncomparing auditory and visual memory retrieval was not the aim of the present study, 396 \nonly auditory-cued picture-recall trials were analyzed. 397 \nAfter the MEG session, we collected participants’ high-resolution structural MRI 398 \ndata for source localization. 399 \n 400 \nMEG data acquisition and preprocessing 401 \nMEG data were recorded using a 275-channel CTF system (VSM MedTech) at a 402 \nsampling rate of 1200 Hz. Preprocessing was performed in Python using the 403 \nMNE-Python toolbox65. Continuous data were band-pass filtered between 1 and 120 404 \nHz using an FIR filter. To remove artifacts, we first applied independent component 405 \nanalysis to the broadband sensor data. Components reflecting eye blinks, saccades, or 406 \ncardiac artifacts were identified through visual inspection of component time courses 407 \nand topographies and then removed. The cleaned data were subsequently epoched 408 \naround stimulus onsets (−0.5 s to 2.5 s) and baseline-corrected using the −0.2 to −0.02 s 409 \npre-stimulus interval. Epochs were down-sampled to 250 Hz and manually inspected to 410 \nreject residual artifacts. 411 \n 412 \nStructural data acquisition and source localization 413 \nStructural MRI data were collected on a Siemens 3T Prisma scanner with a 64-channel 414 \nhead coil. T1-weighted images were acquired using the magnetization-prepared rapid 415 \nacquisition gradient echo (MPRAGE) sequence (TR = 2000 ms, TE = 2.85 ms, field of 416 \nview = 256 × 240 mm, voxel size = 0.8 × 0.8 × 0.8 mm). 417 \nStructural T1-weighted MRIs were processed with FreeSurfer to reconstruct 418 \ncortical surfaces and define the source space66. MEG–MRI coregistration was 419 \nperformed individually, run by run, by aligning fiducial points to anatomical landmarks 420 \n(nasion and bilateral preauricular points). A single-shell boundary element model was 421 \nthen generated to model the inner-skull conductivity, from which individual forward 422 \nsolutions were computed for each experimental run in surface space. Linearly 423 \nconstrained minimum variance spatial filters were constructed for each run using 424 \nempirical data (0.01–2.0 s) and baseline (−0.5 to −0.02 s) covariance matrices. Filters 425 \nwere built with unit-noise-gain normalization and then applied to cleaned MEG epochs 426 \nto obtain single-trial source time courses. Epochs with head displacements exceeding 427 \n10 mm were excluded. Head displacement was computed from the circumcenter of 428 \nthree localization coils per epoch. Source estimates were cropped to −0.4–2.2 s, 429 \nband-pass filtered (1–40 Hz), and down-sampled to 100 Hz. All source data were 430 \ncomputed in native space and parcellated according to the Destrieux cortical atlas 431 \n(aparc.a2009s) for subsequent ROI-based decoding analyses67. 432 \n 433 \nEncoding-retrieval cross-phase classification analysis 434 \nWe quantified reinstatement between perception and retrieval using a ROI-wise, 435 \ntime-resolved cross-phase decoding analysis on source estimates. Based on our 436 \nhypotheses, vmPFC ROIs included left and right G_subcallosal, G_orbital, G_rectus, 437 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 13\nS_suborbital, G_and_S_cingul-Ant, G_and_S_transv_frontopol, 438 \nG_and_S_frontomargin and VVC ROIs included left and right 439 \nG_oc-temp_lat-fusiform, S_oc-temp_lat,S_oc-temp_med_and_Lingual, as defined in 440 \nthe Destrieux (aparc.a2009s) atlas. For each participant, single-trial LCMV source time 441 \ncourses were extracted for each vertex within the ROIs. Within each ROI, we formed 442 \nfeature vectors from vertex-wise source amplitudes at each time point (one vector per 443 \ntrial). Linear support vector machine (SVM) classifiers were trained using data during 444 \nthe encoding phase for each time point. Two kinds of SVM classifiers were trained. 445 \nConceptual classifiers determined whether the picture was a plant or an architecture 446 \npicture, while perceptual classifiers classified whether the picture was a color or 447 \nblack-and-white picture. Then, we examined the conceptual and perceptual 448 \nreinstatement using conceptual and perceptual classifiers, which were tested on 449 \npicture-retrieval trials for each time point. Conceptual and perceptual classification 450 \naccuracy time series were averaged across vmPFC ROIs and VVC ROIs, and time 451 \nseries were temporally smoothed with a Gaussian kernel (FWHM = 25 ms). We 452 \ninterpret higher cross-phase accuracy as stronger reinstatement of category-specific 453 \nperceptual and conceptual representations during picture retrieval. SVM classification 454 \nanalysis was performed using the scikit-learn toolbox in Python68. 455 \nIn the group analysis, we exported conceptual and perceptual reinstatement time 456 \nseries for each participant from the cross-phase classification analysis. The analyzed 457 \ntime window was from 0 to 400 ms. Conceptual and perceptual reinstatement strength 458 \nwere contrasted for each time point within the time window using paired t-tests 459 \nseparately for vmPFC and VVC. Cluster-based permutation testing was performed to 460 \ncontrol for multiple comparisons (cluster-forming p < 0.05; α  = 0.05; 2,000 461 \npermutations). The results were visualized using the seaborn library69. Because we used 462 \nan automatic cued-recall paradigm, participants heard sounds during both encoding and 463 \nretrieval, and these trials served as the training and testing sets for the classification 464 \nanalysis. Importantly, the results are unlikely to reflect auditory perception, because 465 \nsound–picture pairings were randomized across participants. The classification was 466 \ncategorical. Therefore, any given sound item was paired with different perceptual or 467 \nconceptual picture categories across participants. As a result, group-level differences in 468 \nclassification performance reflect differences in perceptual or conceptual picture 469 \ncategory processing rather than sound category processing. 470 \n 471 \nFrequency-specific encoding-retrieval cross-phase classification analysis 472 \nWe filtered the broadband source time series from 2 to 40 Hz in 0.5 Hz steps. Then, the 473 \nsame procedure as the encoding-retrieval cross-phase classification analysis was 474 \nperformed on narrowband time series at each frequency. The averaged vmPFC and 475 \nVVC frequency-specific reinstatement time series were further averaged across four 476 \nfrequency bands: theta (4–8 Hz), alpha (8–13 Hz), beta (13–30 Hz), and gamma (30–40 477 \nHz). The same group analysis comparing conceptual and perceptual reinstatement with 478 \ncluster-based correction was performed for each frequency band. 479 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 14\n 480 \nConceptual and perceptual reinstatement latency comparison 481 \nConceptual and perceptual reinstatement latencies for each participant were defined as 482 \nthe peak latencies within the 0–400 ms range. Paired t-tests were performed to compare 483 \nthe peak latencies between conceptual and perceptual reinstatement. 484 \n 485 \nInformational cross-correlation analysis 486 \nWe performed informational cross-correlation analysis to test lead–lag relationships 487 \nbetween conceptual and perceptual reinstatement. We computed lagged 488 \ncross-correlations on individual ROI-averaged conceptual reinstatement time series in 489 \nthe vmPFC and perceptual reinstatement time series in the VVC, using 15 lags (10–150 490 \nms; with 10 ms steps at 100 Hz). A directionality index D(k) was defined at each lag k 491 \nas the difference between /g1870 /g2913/g2925/g2924/g2913/g2915/g2926/g2930/g2931/g2911/g2922/g1372/g2926/g2915/g2928/g2913/g2915/g2926/g2930/g2931/g2911/g2922/g4666/g1863/g4667  and /g1870 /g2926/g2915/g2928/g2913/g2915/g2926/g2930/g2931/g2911/g2922/g1372/g2913/g2925/g2924/g2913/g2915/g2926/g2930/g2931/g2911/g2922/g4666/g1863/g4667 . 492 \nGroup inference on D(k) used a two-tailed t-test with permutation testing across 493 \nsubjects with cluster-based correction over the lag dimension (cluster-forming p < 0.05; 494 \ncluster α  = 0.05; 2,000 permutations). 495 \n 496 \nSpectral state-space Granger causality analysis 497 \nWe tested directed interactions between vmPFC and VVC during retrieval using 498 \nspectral state-space Granger causality on time-reversed single-trial source estimates 499 \nduring picture retrieval70. Within each ROI, vertex activity was summarized using the 500 \nmne.extract_label_time_course function with “pca_flip” mode. Spectral time-reversed 501 \nGC was computed using vmPFC→ VVC and VVC→ vmPFC index sets 502 \n(mne_connectivity.spectral_connectivity_epochs). For each participant, we obtained 503 \nGC spectra (vmPFC→ VVC and VVC→ vmPFC) across 2 to 40 Hz. Similar to the 504 \nfrequency-specific cross-phase classification analysis, we also averaged GC estimates 505 \nacross four frequency bands: theta (4–8 Hz), alpha (8–13 Hz), beta (13–30 Hz), and 506 \ngamma (30–40 Hz). Then, we performed a paired t-test between GC (vmPFC→ VVC) 507 \nand GC (VVC→ vmPFC) to examine whether the information flow during picture 508 \nretrieval is from vmPFC to VVC or vice versa. 509 \n 510 \n  511 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 15\nAcknowledgments 512 \nFunding: This work was supported by  513 \nCompeting interests: The authors declare that they have no competing financial 514 \ninterests. 515 \nData and materials availability: The paper and/or the supplementary materials 516 \ncontain all the data needed to evaluate the conclusions. The behavioral and fMRI data 517 \nthat support this study's findings will be available on OSF. 518 \n 519 \n  520 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 16\nReference 521 \n1. V ogt, S. & Magnussen, S. Expertise in Pictorial Perception: Eye-Movement Patterns and 522 \nVisual Memory in Artists and Laymen. Perception 36, 91–100 (2007). 523 \n2. Wing, E. A., Burles, F., Ryan, J. D. & Gilboa, A. The structure of prior knowledge enhances 524 \nmemory in experts by reducing interference. Proc. Natl. Acad. Sci. U.S.A.  119, e2204172119 525 \n(2022). 526 \n3. Konkle, T., Brady, T. F., Alvarez, G. A. & Oliva, A. Conceptual distinctiveness supports 527 \ndetailed visual long-term memory for real-world objects. Journal of Experimental Psychology: 528 \nGeneral 139, 558–578 (2010). 529 \n4. Josselyn, S. A., Köhler, S. & Frankland, P. W. Finding the engram. Nat Rev Neurosci 16, 530 \n521–534 (2015). 531 \n5. Josselyn, S. A. & Tonegawa, S. Memory engrams: Recalling the past and imagining the future. 532 \nScience 367, eaaw4325 (2020). 533 \n6. Teyler, T. J. & DiScenna, P. The hippocampal memory indexing theory. Behavioral 534 \nNeuroscience 100, 147–154 (1986). 535 \n7. Hebscher, M., Wing, E., Ryan, J. & Gilboa, A. Rapid Cortical Plasticity Supports Long-Term 536 \nMemory Formation. Trends in Cognitive Sciences 23, 989–1002 (2019). 537 \n8. Moscovitch, M., Cabeza, R., Winocur, G. & Nadel, L. Episodic Memory and Beyond: The 538 \nHippocampus and Neocortex in Transformation. Annu. Rev. Psychol. 67, 105–134 (2016). 539 \n9. Sekeres, M. J., Winocur, G. & Moscovitch, M. The hippocampus and related neocortical 540 \nstructures in memory transformation. Neuroscience Letters 680, 39–53 (2018). 541 \n10. Mesulam, M. M. From sensation to cognition. Brain 121, 1013–1052 (1998). 542 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 17\n11. Ranganath, C. & Ritchey, M. Two cortical systems for memory-guided behaviour. Nat Rev 543 \nNeurosci 13, 713–726 (2012). 544 \n12. Bone, M. B., Ahmad, F. & Buchsbaum, B. R. Feature-specific neural reactivation during 545 \nepisodic memory. Nat Commun 11, 1945 (2020). 546 \n13. Dijkstra, N., Bosch, S. E. & Gerven, M. A. J. van. Vividness of Visual Imagery Depends on 547 \nthe Neural Overlap with Perception in Visual Areas. J. Neurosci. 37, 1367–1373 (2017). 548 \n14. Xue, G. The Neural Representations Underlying Human Episodic Memory. Trends in 549 \nCognitive Sciences 22, 544–561 (2018). 550 \n15. Wheeler, M. E., Petersen, S. E. & Buckner, R. L. Memory’s echo: Vivid remembering 551 \nreactivates sensory-specific cortex. Proceedings of the National Academy of Sciences  97, 552 \n11125–11129 (2000). 553 \n16. Ritchey, M., Wing, E. A., LaBar, K. S. & Cabeza, R. Neural Similarity Between Encoding 554 \nand Retrieval is Related to Memory Via Hippocampal Interactions. Cerebral Cortex  23, 555 \n2818–2828 (2013). 556 \n17. Bartlett, F. C. & Kintsch, W. Remembering: A Study in Experimental and Social Psychology . 557 \n(Cambridge University Press, 1995). doi:10.1017/CBO9780511759185. 558 \n18. Schacter, D. L. & Addis, D. R. The cognitive neuroscience of constructive memory: 559 \nremembering the past and imagining the future. Philosophical Transactions of the Royal 560 \nSociety B: Biological Sciences 362, 773–786 (2007). 561 \n19. Schacter, D. L., Norman, K. A. & Koutstaal, W. The cognitive neuroscience of constructive 562 \nmemory. in False-memory creation in children and adults:  Theory, research, and 563 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 18\nimplications 129–168 (Lawrence Erlbaum Associates Publishers, Mahwah, NJ, US, 2000). 564 \n20. Zhang, L., Alain, C. & Buchsbaum, B. R. Differential reliance on sensory reinstatement and 565 \ninternally transformed representation during vivid retrieval of visual and auditory episodes. 566 \nPreprint at https://doi.org/10.1101/2025.06.16.659408 (2025). 567 \n21. Gordon, A. M., Rissman, J., Kiani, R. & Wagner, A. D. Cortical Reinstatement Mediates the 568 \nRelationship Between Content-Specific Encoding Activity and Subsequent Recollection 569 \nDecisions. Cereb Cortex 24, 3350–3364 (2014). 570 \n22. Barker, R. M., St-Laurent, M. & Buchsbaum, B. R. Neural reactivation and judgements of 571 \nvividness reveal separable contributions to mnemonic representation. NeuroImage  255, 572 \n119205 (2022). 573 \n23. Heinbockel, H., Wagner, A. D. & Schwabe, L. Post-retrieval stress impairs subsequent 574 \nmemory depending on hippocampal memory trace reinstatement during reactivation. Science 575 \nAdvances 10, eadm7504 (2024). 576 \n24. Bowman, C. R. & Zeithamova, D. Abstract Memory Representations in the Ventromedial 577 \nPrefrontal Cortex and Hippocampus Support Concept Generalization. J. Neurosci.  38, 578 \n2605–2614 (2018). 579 \n25. Ghosh, V. E., Moscovitch, M., Colella, B. M. & Gilboa, A. Schema Representation in Patients 580 \nwith Ventromedial PFC Lesions. J. Neurosci. 34, 12057–12070 (2014). 581 \n26. Giuliano, A. E., Bonasia, K., Ghosh, V . E., Moscovitch, M. & Gilboa, A. Differential 582 \nInfluence of Ventromedial Prefrontal Cortex Lesions on Neural Representations of Schema 583 \nand Semantic Category Knowledge. J Cogn Neurosci 33, 1928–1955 (2021). 584 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 19\n27. Kerrén, C., Reznik, D., Doeller, C. F. & Griffiths, B. J. Exploring the role of dimensionality 585 \ntransformation in episodic memory. Trends in Cognitive Sciences 29, 614–626 (2025). 586 \n28. Mack, M. L., Preston, A. R. & Love, B. C. Ventromedial prefrontal cortex compression 587 \nduring concept learning. Nat Commun 11, 46 (2020). 588 \n29. Bein, O. & Niv, Y . Schemas, reinforcement learning and the medial prefrontal cortex. Nat. 589 \nRev. Neurosci. 26, 141–157 (2025). 590 \n30. Lifanov, J., Linde-Domingo, J. & Wimber, M. Feature-specific reaction times reveal a 591 \nsemanticisation of memories over time and with repeated remembering. Nat Commun  12, 592 \n3177 (2021). 593 \n31. Linde-Domingo, J., Treder, M. S., Kerrén, C. & Wimber, M. Evidence that neural information 594 \nflow is reversed between object perception and object reconstruction from memory. Nat 595 \nCommun 10, 179 (2019). 596 \n32. Guth, T. A. et al.  Theta-phase locking of single neurons during human spatial memory. Nat 597 \nCommun 16, 7402 (2025). 598 \n33. Herweg, N. A., Solomon, E. A. & Kahana, M. J. Theta Oscillations in Human Memory. 599 \nTrends in Cognitive Sciences 24, 208–227 (2020). 600 \n34. Kota, S., Rugg, M. D. & Lega, B. C. Hippocampal Theta Oscillations Support Successful 601 \nAssociative Memory Formation. J. Neurosci. 40, 9507–9518 (2020). 602 \n35. Rudoler, J. H., Herweg, N. A. & Kahana, M. J. Hippocampal Theta and Episodic Memory. J. 603 \nNeurosci. 43, 613–620 (2023). 604 \n36. Ursino, M. & Pirazzini, G. Theta–gamma coupling as a ubiquitous brain mechanism: 605 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 20\nimplications for memory, attention, dreaming, imagination, and consciousness. Current 606 \nOpinion in Behavioral Sciences 59, 101433 (2024). 607 \n37. Anderson, K. L., Rajagovindan, R., Ghacibeh, G. A., Meador, K. J. & Ding, M. Theta 608 \nOscillations Mediate Interaction between Prefrontal Cortex and Medial Temporal Lobe in 609 \nHuman Memory. Cereb Cortex 20, 1604–1612 (2010). 610 \n38. Backus, A. R., Schoffelen, J.-M., Szebényi, S., Hanslmayr, S. & Doeller, C. F. 611 \nHippocampal-Prefrontal Theta Oscillations Support Memory Integration. Current Biology 26, 612 \n450–457 (2016). 613 \n39. Hebscher, M., Meltzer, J. A. & Gilboa, A. A causal role for the precuneus in network-wide 614 \ntheta and gamma oscillatory activity during complex memory retrieval. eLife 8, e43114 615 \n(2019). 616 \n4 0. B ar r y , D. N., B ar n e s , G . R. , C l ar k , I . A . & Ma gu i r e , E. A . T h e Ne ur al  Dy na mi cs  of  Nov e l  617 \nScene Imagery. J. Neurosci. 39, 4375–4386 (2019). 618 \n41. Fuentemilla, L., Barnes, G. R., Düzel, E. & Levine, B. Theta oscillations orchestrate medial 619 \ntemporal lobe and neocortex in remembering autobiographical memories. NeuroImage 85, 620 \n730–737 (2014). 621 \n42. Kaplan, R. et al.  Medial prefrontal theta phase coupling during spatial memory retrieval. 622 \nHippocampus 24, 656–665 (2014). 623 \n43. Barron, H. C. et al.  Unmasking Latent Inhibitory Connections in Human Cortex to Reveal 624 \nDormant Cortical Memories. Neuron 90, 191–203 (2016). 625 \n44. Tomita, H., Ohbayashi, M., Nakahara, K., Hasegawa, I. & Miyashita, Y. Top-down signal 626 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 21\nfrom prefrontal cortex in executive control of memory retrieval. Nature 401, 699–703 (1999). 627 \n45. Guise, K. G. & Shapiro, M. L. Medial Prefrontal Cortex Reduces Memory Interference by 628 \nModifying Hippocampal Encoding. Neuron 94, 183-192.e8 (2017). 629 \n46. McCormick, C., Barry, D. N., Jafarian, A., Barnes, G. R. & Maguire, E. A. vmPFC Drives 630 \nHippocampal Processing during Autobiographical Memory Recall Regardless of Remoteness. 631 \nCereb Cortex 30, 5972–5987 (2020). 632 \n47. Robin, J. & Moscovitch, M. Details, gist and schema: Hippocampal–neocortical interactions 633 \nunderlying recent and remote episodic and spatial memory. Current Opinion in Behavioral 634 \nSciences 17, 114–123 (2017). 635 \n48. Mesulam, M.-M. Behavioral Neuroanatomy Large-Scale Networks, Association Cortex, 636 \nFrontal Syndromes, the Limbic System, and Hemispheric Specializations. in Principles of 637 \nBehavioral and Cogn itive Neurology  (ed. Mesulam, M.-M.) 0 (Oxford University Press, 638 \n2000). doi:10.1093/oso/9780195134759.003.0001. 639 \n49. Winocur, G. & Moscovitch, M. Memory Transformation and Systems Consolidation. Journal 640 \nof the International Neuropsychological Society 17, 766–780 (2011). 641 \n50. Favila, S. E., Lee, H. & Kuhl, B. A. Transforming the Concept of Memory Reactivation. 642 \nTrends in Neurosciences 43, 939–950 (2020). 643 \n51. Gloede, M. E. & Gregg, M. K. The fidelity of visual and auditory memory. Psychon Bull Rev 644 \n26, 1325–1332 (2019). 645 \n52. Clarke, A., Taylor, K. I., Devereux, B., Randall, B. & Tyler, L. K. From Perception to 646 \nConception: How Meaningful Objects Are Processed over Time. Cereb Cortex 23, 187–197 647 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 22\n(2013). 648 \n53. Gilboa, A. & Moscovitch, M. Ventromedial prefrontal cortex generates pre-stimulus theta 649 \ncoherence desynchronization: A schema instantiation hypothesis. Cortex 87, 16–30 (2017). 650 \n54. Griffiths, B. J. & Jensen, O. Gamma oscillations and episodic memory. Trends in 651 \nNeurosciences 46, 832–846 (2023). 652 \n55. Mysin, I. & Shubina, L. From mechanisms to functions: The role of theta and gamma 653 \ncoherence in the intrahippocampal circuits. Hippocampus 32, 342–358 (2022). 654 \n56. Min Park, Y ., Park, J., Young Kim, I., Koo Kang, J. & Pyo Jang, D. Interhemispheric theta 655 \ncoherence in the hippocampus for successful object-location memory in human intracranial 656 \nencephalography. Neuroscience Letters 786, 136769 (2022). 657 \n57. Nyhus, E. & Curran, T. Functional role of gamma and theta oscillations in episodic memory. 658 \nNeuroscience & Biobehavioral Reviews 34, 1023–1035 (2010). 659 \n58. Welle, C. G. & Contreras, D. Sensory-driven and spontaneous gamma oscillations engage 660 \ndistinct cortical circuitry. J Neurophysiol 115, 1821–1835 (2016). 661 \n59. Wilming, N., Murphy, P. R., Meyniel, F. & Donner, T. H. Large-scale dynamics of perceptual 662 \ndecision information across human cortex. Nat Commun 11, 5109 (2020). 663 \n60. Lizarazu, M., Lallier, M. & Molinaro, N. Phase−amplitude coupling between theta and 664 \ngamma oscillations adapts to speech rate. Annals of the New York Academy of Sciences  1453, 665 \n140–152 (2019). 666 \n61. Busch, N. A., Dubois, J. & VanRullen, R. The Phase of Ongoing EEG Oscillations Predicts 667 \nVisual Perception. J. Neurosci. 29, 7869–7876 (2009). 668 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 23\n62. Michalareas, G. et al. Alpha-Beta and Gamma Rhythms Subserve Feedback and Feedforward 669 \nInfluences among Human Visual Cortical Areas. Neuron 89, 384–397 (2016). 670 \n63. Bastos, A. M. et al.  Canonical Microcircuits for Predictive Coding. Neuron 76, 695–711 671 \n(2012). 672 \n64. Faul, F., Erdfelder, E., Lang, A.-G. & Buchner, A. G*Power 3: a flexible statistical power 673 \nanalysis program for the social, behavioral, and biomedical sciences. Behav Res Methods  39, 674 \n175–191 (2007). 675 \n65. Larson, E. et al. MNE-Python. Zenodo https://doi.org/10.5281/ZENODO.592483 (2025). 676 \n66. Fischl, B. FreeSurfer. NeuroImage 62, 774–781 (2012). 677 \n67. Destrieux, C., Fischl, B., Dale, A. & Halgren, E. Automatic parcellation of human cortical 678 \ngyri and sulci using standard anatomical nomenclature. NeuroImage 53, 1–15 (2010). 679 \n68. Pedregosa, F. et al.  Scikit-learn: Machine Learning in Python. J. Mach. Learn. Res.  12, 680 \n2825–2830 (2011). 681 \n69. Waskom, M. seaborn: statistical data visualization. JOSS 6, 3021 (2021). 682 \n70. Winkler, I., Panknin, D., Bartz, D., Muller, K.-R. & Haufe, S. Validity of Time Reversal for 683 \nTesting Granger Causality. IEEE Trans. Signal Process. 64, 2746–2760 (2016). 684 \n 685 \n 686 \n 687 \n  688 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 24\n 689 \nFigure 1. Experimental paradigm and analysis pipeline. (A) Four picture stimuli 690 \nused in the experiment were organized along two orthogonal dimensions: a perceptual 691 \ndimension (color vs. black-and-white) and a conceptual dimension (architecture vs. 692 \nplant). (B) MEG run structure. Each run began with twenty-four encoding trials in 693 \nwhich participants viewed intact audiovisual pairs. This was followed by twenty-four 694 \ncued-recall trials, including auditory-cued picture retrieval and picture-cued sound 695 \nretrieval. (C) After MEG data acquisition, sensor-level data were source-localized 696 \nusing an LCMV beamformer. We then extracted trial-wise time series from vmPFC and 697 \nVVC regions of interest. Encoding–retrieval MVPA was performed at each time point 698 \nto quantify perceptual and conceptual reinstatement. vmPFC, ventromedial prefrontal 699 \ncortex; VVC, ventral visual cortex; MVPA, multivariate pattern analysis. 700 \n 701 \n 702 \n 703 \n  704 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 25\n 705 \nFigure 2. Conceptual reinstatement in vmPFC precedes perceptual reinstatement 706 \nin VVC. (A). vmPFC showed stronger conceptual reinstatement than perceptual 707 \nreinstatement from 110-160 ms (/g1868 /g3033/g3050/g3032 < 0.05). (B) VVC showed stronger perceptual 708 \nreinstatement than conceptual feature reinstatement from 190-230 ms (/g1868 /g3033/g3050/g3032 < 0.05). 709 \n(C) The onset latency of conceptual reinstatement in vmPFC was significantly earlier 710 \nthan the onset of perceptual reinstatement in VVC (/g1868 /g3033/g3050/g3032 < 0.05). vmPFC, 711 \nventromedial prefrontal cortex; VVC, ventral visual cortex; **, p < 0.01. 712 \n 713 \n  714 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 26\n 715 \nFigure 3. Conceptual advantage in vmPFC expressed in theta band while leads 716 \nperceptual advantage in VVC expressed in gamma band. (A). In vmPFC, 717 \nconceptual reinstatement was significantly stronger than perceptual reinstatement in 718 \nthe theta band during the two time windows: 0–70 ms and 110–160 ms (  < 0.05). 719 \n(B) In contrast, VVC showed significantly stronger perceptual than conceptual 720 \nreinstatement in the gamma band from 300–370 ms (  < 0.05). vmPFC, 721 \nventromedial prefrontal cortex; VVC, ventral visual cortex. 722 \n 723 \n 724 \n 725 \n  726 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 27\n 727 \nFigure 4. Conceptual reinstatement in vmPFC directly influenced perceptual 728 \nreinstatement in VVC. (A) vmPFC conceptual reinstatement exerted a direct 729 \ninfluence on VVC perceptual reinstatement during the 70–90 ms time window. (B). 730 \nStronger conceptual reinstatement could significantly predict later perceptual 731 \nreinstatement. vmPFC, ventromedial prefrontal cortex; VVC, ventral visual cortex. 732 \n 733 \n  734 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 28\n 735 \nFigure 5. Bi-directional connectivity between vmPFC and VVC. vmPFC drives 736 \nVVC activity via theta-band top-down influences during retrieval, whereas VVC 737 \nprovides bottom-up feedback to vmPFC through alpha-band connectivity. vmPFC, 738 \nventromedial prefrontal cortex; VVC, ventral visual cortex; *, p < 0.05; **, p < 0.01. 739 \n 740 \n 741 \n  742 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint \n\n 29\n 743 \nFigure 6. Hierarchical interactive model of episodic memory retrieval. During 744 \nepisodic retrieval, conceptual information is first reactivated in vmPFC, which provides 745 \nhigh-level, schema-like constraints. This conceptual signal guides and constrains 746 \nperceptual reinstatement in VVC via top-down theta-band oscillatory interactions. 747 \nMeanwhile, VVC sends bottom-up alpha-band feedback to vmPFC, supplying 748 \nperceptual evidence for further evaluation and integration. Together, these bidirectional 749 \ninteractions support the hierarchical reconstruction of vivid episodic memories. 750 \nvmPFC, ventromedial prefrontal cortex; VVC, ventral visual cortex. 751 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted December 20, 2025. ; https://doi.org/10.64898/2025.12.17.695049doi: bioRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}