Biology subjects

Gozukara, D.

Publications and source records attributed to Gozukara, D..

3 recordsLinked to original sources

bioRxiv2026

Memory and Hippocampal Responses to Event Boundaries are Modulated by Global Brain States

Our daily experiences unfold as a continuous stream, yet we perceive and remember them as discrete events. Event boundaries, the moments of transition between these events, are known to elicit increases in hippocampal activity believed to reflect memory encoding. However, it remains unknown how this hippocampal response relates to large-scale brain dynamics. Here, using fMRI data from two independent datasets (Sherlock and StudyFor-rest), we applied the Greedy State Boundary Search (GSBS) algorithm to whole-brain activity patterns and identified two recurring global brain states corresponding to the Default Mode Network (DMN) and Task-Positive Network (TPN). We found that event boundaries were associated with an increased probability of being in the TPN state, and that hippocampal activity was generally higher during TPN states. The hippocampal response to event boundaries appeared predominantly during TPN states. When overall state-related differences in baseline hippocampal activity were controlled for, event boundaries elicited a hippocampal response regardless of the concurrent global state. Critically, individual differences in the tendency to shift toward the TPN state at event boundaries; as well as overall time spent at the DMN state predicted subsequent memory for narrative content, whereas univariate hippocampal activity at boundaries did not. These findings demonstrate that hippocampal event boundary responses are modulated by global brain state dynamics, and suggest that the interplay between large-scale network configurations and event segmentation plays a key role in how continuous experience is encoded into memory.

neuroscience↗

bioRxiv2025

Multi-Scale Anti-Correlated Neural States Dominate Naturalistic Whole-Brain Activity

The human brains response to naturalistic stimuli is characterized by complex spatiotemporal dynamics. Within these dynamics there is a transitioning structure between sets of anti-correlated neural states that is frequently observed but has not been systematically investigated across scales. In this paper, we use three different naturalistic fMRI datasets to quantify anti-correlation in global and local neural states during naturalistic viewing or listening and investigate their interdependence and their relationship to changes in the stimuli. We demonstrate that continuous naturalistic brain activity shows an anti-correlational structure that spans both global and local spatial scales, with regions in the dorsal attention network showing strong alignment between local and global state transitions. On the global scale, ongoing dynamics are dominated by two antagonistic states that correspond to Default Mode Network and Task Positive Network configurations, with a third transitional state mediating between them. On the local scale, we observe anti-correlated neural states that are associated with periods of relatively high and low brain activity. Across the brain, these are driven by subsets of voxels that are systematically anti-correlated with their areas dominant activity pattern. This antagonism is related to stimulus changes, which tend to trigger a switch to the TPN state globally and to high activity states locally. On the local scale we also see a modality-specific pattern, with visual changes mostly driving transitions in visual cortical regions and auditory changes predominantly affecting auditory and language-related areas. The consistency of these findings across datasets with different stimulus types (audiovisual and purely auditory) indicates that anti-correlated neural states represent a domain-general organizational principle of brain function. We propose that anti-correlated dynamics functionally represent a convergent solution to the fundamental challenge of maintaining coherent internal representations while remaining responsive to meaningful changes in the environment.

neuroscience↗

bioRxiv2025

Neural network-based encoding in free-viewing fMRI with precision models

Representations learned by convolutional neural networks (CNNs) exhibit a remarkable resemblance to information processing patterns observed in the primate visual system on large neuroimaging datasets collected under diverse, naturalistic visual stimulation, but with instruction for participants to maintain central fixation. This viewing condition, however, diverges significantly from ecologically valid visual behaviour, suppresses activity in visually active regions, and imposes substantial cognitive load on the viewing task. We present a modification of the encoding model framework, adapting it for use with naturalistic vision datasets acquired under fully natural viewing conditions, without fixation, by incorporating eye-tracking data. Our gaze-aware encoding models were trained on the StudyForrest dataset, which features task-free naturalistic movie viewing. By combining eye-tracking data with the visual content of movie frames, we generate combined subject-wise gaze-stimulus specific feature time series. These time series are constructed by sampling only the locally and temporally relevant elements of the CNN feature map for each fixation. Our results demonstrate that gaze-aware encoding models match the performance of conventional encoding models with 112x fewer model parameters. Gaze-aware encoding models were especially beneficial for participants with more dynamic eye-movement patterns. Therefore, this approach opens the door to more ecologically valid models that can be built in more naturalistic settings, such as playing games or navigating virtual environments.

neuroscience↗