Investigating the neural mechanisms of human cognitive function through intracranial recordings

National Institute Of Neurological Disorders And Stroke

Investigators

Linked publications, trials & patents

Abstract

FY2023 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. To this end, we have uncovered a number of important phenomena. In particular, we have found that spiking activity in the human anterior temporal lobe cortex is often organized into bursts of spiking activity, and these bursts of neuronal firing are organized into temporal sequences. Our recent efforts have focused on understanding how these sequences may be relevant for neural coding. Our hypothesis has been that this sequential order of spiking activity may be a fundamental building block for representing information in the brain. By examining these sequences of spiking activity as participants view images drawn from different semantic categories, we have now 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. We have recently submitted a manuscript for publication describing this work in detail (Wittig JH, Xie W, Jackson SJ, Inati SK, Zaghloul KA (2023) In Review). This finding aligns with our previous work showing that the order of spiking activity that is present as individuals encode items into their memory is replayed when they retrieve those same items, suggesting that the order of spiking activity carries specific information about the items we are representing in our brains. We are now linking these findings with another phenomenon that we have recently described in our work related to 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. We have previously found that small local patches of cortex, which we refer to as modules, are highly connected and that activity within these modules is differentially modulated by different stimuli, suggesting that these modules encode different functional information. These modules exhibit the same spatial dimensions and functional characteristics of cortical columns that are hypothesized to exist throughout the human cortex. Examining the sequences of spiking activity, we have now found that each sequence is composed of smaller sub-sequences that arise from the activity of neurons in the different cortical modules that we have identified. During any given burst of spiking activity, the neurons in some but not all modules within a patch of cortex are actively involved. Hence, any one individual sequence is in fact composed of spiking activity drawn in combinatorial manner from the spiking activity in different cortical modules. We are currently preparing a manuscript describing this work (Chapeton J, Swift K, Wittig JH, Inati SK, Zaghloul KA (2022) 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 completed a study in which we identified that the specific sequences of spiking activity that we observe in our data are both specific to individual items being represented in the brain but also are relatively constrained in their overall order. The spiking sequences observed in an individual patch of cortex during different times are, in general, somewhat similar to one another. Our data therefore suggest that the spiking activity of any patch of cortex can be characterized by a relatively stable backbone of sequential firing, around which individual variations in spiking order may enable the representations of item-specific information. We have recently published a manuscript describing this work (Vaz AP, Wittig JH, Inati SK, Zaghloul KA (2023) Nature Communications). In another study, we have also examined how network interactions at a larger spatial scale may be relevant for memory encoding and retrieval. We specifically examined large scale connectivity in the brain and identified pairs of regions that demonstrate stable time delays in their correlated activity. The presence of a reliable and consistent time delay suggests that indeed two brain regions may have a reliable connection. Based on these pairs alone, we then examined how the moment-to-moment fluctuations in connectivity form a unique pattern of connectivity for every item that we remember, and how these patterns are reinstated when retrieving these memories. These changes in connectivity that form a neural signature of the item being remembered are dissociable from changes in spectral power, suggesting that connectivity itself contains information about the events we encode into memory. We have completed this work and recently submitted a manuscript describing this work for publication (Phan AP, Xie W, Chapeton J, Inati SK, Zaghloul KA (2023) In Review). We have also examined how long-term episodic memory may interact with short-term memory (STM). There are significant parallels between the computations required for the precision of STM and those required to minimize mnemonic interference for long-term memory, a process referred to as pattern separation that has traditionally been ascribed to the medial temporal lobe (MTL). We have recently uncovered direct evidence supporting the hypothesis that the circuitry of the MTL, long viewed as a dedicated module for distinguishing similar neural representations of information in long-term memory, also supports STM precision. We recently published a manuscript describing this work (Xie W, Wittig JH, Bhasin S, Zawora C, Inati SK, Zhang W, Zaghloul KA (2022) Nature Human Behaviour). Finally, we have also spent much effort over the past year examining how these recordings may provide us some insight into the basic mechanisms of epilepsy, and how seizures begin and propagate throughout the brain. We have found that discharges appear to observe a stereotyped and constrained pattern of propagation. Because of this constraint, we can use the pattern of propagation to estimate the pathological source of the discharges. We have recently published a manuscript describing this work in detail (Diamond JM, Withers CP, Chapeton JI, Rahman S, Inati SK, Zaghloul KA (2023) Brain). We have also examined the role of direct electrical stimulation in modulating neural activity. We routinely use stimulation for clinical purposes, and one of our long-term goals is to use direct stimulation to investigate questions of causality. We have found that neurons can be selectively excited or inhibited by microstimulation, and that different stimulation sites may have different effects on the same neurons. We have completed this work and recently published a manuscript describing this study (Yousef D, Wittig JH, Inati SK, Zaghloul KA (2023) J Neuroscience).