Investigating the neural mechanisms of human cognitive function through intracranial recordings

National Institute Of Neurological Disorders And Stroke

Investigators

Linked publications & trials

Abstract

FY2025 has seen significant progress towards realizing our goals and objectives. We have continued our efforts capturing and analyzing intracranial recordings while participants engage in cognitive tasks designed to probe memory encoding and retrieval. Patients with drug resistant epilepsy receiving intracranial electrodes and surgical treatment at the Clinical Center have been recruited for these studies. Our work takes advantage of the opportunities to record both intracranial EEG and single unit spiking activity from these implanted electrodes as participants perform a variety of cognitive tasks during the monitoring period. Our efforts are focused on understanding changes in human brain activity across these different spatial scales. In order to probe how memories are formed and retrieved in the brain, we have focused our efforts on understanding how information is represented in the neural signals that we record. In particular, over the past year we published a study in which we demonstrated that spiking activity in the human anterior temporal lobe cortex is often organized into bursts of spiking activity that are organized into specific temporal sequences (Xie W, Wittig JH, Chapeton JI, Jackson SJ, Inati SK, Zaghloul KA (2024) Nature). We found direct evidence that the specific order of neuronal firing in the anterior temporal lobe can distinguish different semantic categories, and therefore temporal order of neural activity seems to play a role in representing different semantic information. Building upon this, we have spent significant efforts over the past year examining how connectivity at the smallest spatial scales may provide insights into the functional organization and structure of the human cortex, and how such organization may be relevant for encoding information. Specifically, leveraging the high spatial resolution of microelectrode recordings, we have identified individual cortical modules, approximately 1mm in diameter that match the hypothesized diameter of a cortical column. These modules are highly connected small patches of cortex. During any given burst of spiking activity, the neurons in some but not all modules within a patch of cortex are actively involved. We have found that different modules encode different semantic features, and individual neurons within each module match the encoding characteristics of the module to which it belongs. Activity within these modules is differentially modulated by different stimuli, suggesting that these modules encode different functional information. Moreover, the representation of information appears hierarchical – modules can encode the general category of a stimulus but individual neurons are required to represent the individual exemplars. We are finalizing analyses related to this work and are beginning preparation of a manuscript describing this work (Chapeton J, Swift K, Wittig JH, Inati SK, Zaghloul KA (2025) In Preparation). Given these fundamental insights into how neural activity may be organized across these different spatial scales, we have also spent much of our effort over the past year building upon this foundation in order to gain better insights into the neural mechanisms that underlie human episodic memory formation. We have recently published a study in which we have been investigating how the fidelity of these spiking sequences may contribute to memory encoding (Sundby K, Wittig JH, Inati SK, Zaghloul KA (2024) Current Biology). We know from previous work in our lab that there is a strong relation between success memory encoding and attention. In this study, we therefore examined how attention may shape these spiking sequences and found that indeed, when individuals are paying attention, the consistency of the observed spiking sequences increases. In a related study, we have also examined how the fidelity of these sequences may be shaped by whether participants are actively encoding items into memory. We have found that indeed, these sequences convey information regarding semantic category only in cases when participants are actively attending to and attempting to remember images in memory. We have completed a manuscript describing this work and are currently in the process of revising our manuscript for publication (Jitendran E, Zhang J, Wittig JH, Jackson S, Inati SK, Behrens TEJ, Zaghloul KA (2025) In Revision). In another study, we have examined how time and temporal context may be represented by neural signals in the brain. One of the key features of episodic memory is that every event or episode we remember is characterized by the specific time and place where that event occurred. Thus, how the brain represents time plays a crucial role in our ability to accurately form and retrieve memories. To this end, by examining our single unit population spiking data, we have found that the overall spiking representations exhibit a slow temporal drift but also convey information about the exact timing of the cognitive task. For example, we can decode whether a participant is studying the first or second word in a list and how much time has elapsed since a word came on the screen. Interestingly, these slow temporal drifts are interrupted by periodic bursts of spiking, as we have seen in our other data, that do not contain any timing information, suggesting that they may represent some internal aperiodic process used for processing information. We are currently completing analyses for this project and are preparing a manuscript describing this work for submission (Bright I, Swift K, Fruchet O, Jackson S, Howard M, Zaghloul KA (2025) In Preparation. Finally, we have also spent much effort over the past year examining the causal nature of these signals. In particular, we would like to understand whether these neural biomarkers of memory play a causal role in successful encoding and retrieval. One of the primary tools that we can use for testing causality is direct electrical stimulation. As such, we have initiated a number of studies over the past several years in which we have explored the effects of direct electrical stimulation on neural activity, and consequently whether we can use stimulation to evoke particular neural signals. Over the past year, we have completed a study in which we generate computational models characterizing the effects of stimulation, and assess the extent to which those models can accurately predict the outcomes of stimulation. We are currently completing analyses for this project and are preparing a manuscript describing this work for submission (Mohan U, Fruchet O, Jackson S, Zaghloul KA (2025) In Preparation). We have also found that using microstimulation, we can reliable evoke high frequency ripple oscillations in the medial temporal lobe and in the temporal lobe cortex. Previous work in our lab has demonstrated that ripple activity is highly correlated with successful memory formation. Based on this work, we have now initiated a study in which we explore whether evoking ripples precisely locked to the timing of memory encoding and retrieval may affect the fidelity of memory encoding. Data collection and analyses for this project are ongoing.