A neural code for spatiotemporal context

Abstract

Time and space are principle organizing dimensions of human experience. Whereas separate lines of study have identified neural correlates of time and space, little is known about how these representations converge during self-guided experience. Here we asked how neurons in the human brain represent time and space concurrently. Subjects fitted with intracranial microelectrodes played a timed navigation game where they alternated between searching for and retrieving objects in a virtual environment. Significant proportions of both time- and place-selective neurons were present during navigation, and distinct time-selective neurons appeared during task-free delays absent movement. We find that temporal and spatial codes are dissociable, with time cells remapping between search and retrieval tasks while place cells maintained stable firing fields. Other neurons tracked the context unique to each task phase, independent of time or space. Together these neuronal classes comprise a biological basis for the cognitive map of spatiotemporal context.

The influence of emotion on temporal context models

“…The standard conception of temporal context has properties that limit the applicability of computational models of list-learning experiments to real-world memory phenomena. These alternative mathematical forms for temporal context are consistent with a broad range of neurophysiological findings from animals (Bright et al, 2020;MacDonald, Lepage, Eden, & Eichenbaum, 2011;Tsao et al, 2018) and humans (Schonhaut, Aghajan, Kahana, & Fried, 2022). Recordings from populations of neurons have confirmed quantitative predictions of these equations (Cao, Bladon, Charczynski, Hasselmo, & Howard, 2022).…”

Constructing Condensed Memories in Functorial Time

“…Indeed, it is the intuitive nature of space-time independence that renders relativistic physics counter-intuitive. Recent experimental work indicates that space and time representations are dissociable at the functional-network level in humans for short-duration event sequences (Schonhaut, Aghajan, Kahana & Fried, 2022), suggesting at least that the same is true for the long-time representations of interest here. Such dissociability is assumed by all models that treat spatial representations as "maps" that are invariant across time, as they must be to support flexible composition of sequences of actions either in memories or in plans (McNamee et al, 2022).…”

MTL neurons phase-lock to human hippocampal theta

“…During each session, subjects played one of several first-person navigation games in which they freely explored a virtual environment and retrieved objects or navigated to specific locations. The details of these experiments have been previously described ( Ekstrom et al, 2003 ; Jacobs et al, 2010 ; Schonhaut et al, 2022 ); for the purposes of the present study, we pooled data across these studies to generate a large sample for conducting electrophysiological analyses. We analyzed intervals in which subjects could freely navigate through the virtual environment.…”

“…We performed semi-automatic spike sorting and quality inspection on each microwire channel using the WaveClus software package in Matlab ( Quiroga et al, 2004 ), as previously described ( Ekstrom et al, 2003 ; Schonhaut et al, 2022 ). We isolated 0-8 units on each microwire channel, retaining both single-units and multi-units for subsequent analysis while removing units with low amplitude waveforms relative to the noise floor, non-neuronal shapes, inconsistent firing across the recording session, or other data quality issues.…”