| Invited Symposia | Symposia

Note: Symposia have not yet been scheduled.

The rhythms of recollection in the human hippocampus

Chair: L Joel Voss¹; ¹University of Chicago
Presenters: Josh Jacobs, Brad Lega, Anli Liu, Joel Voss

This symposium unites four lines of intracranial human research to examine how hippocampal oscillations organize memory across spatial, temporal, semantic, and behavioral dimensions. Memory is coordinated with medial-temporal-lobe theta and ripple oscillations that are present during encoding of content (objects, people, events) with context (temporal, spatial) and during recollection. Josh Jacobs will describe findings from direct brain recordings during navigation and episodic memory, indicating that human place- and grid-like coding extends beyond the classic hippocampal-entorhinal system and may be organized by broader traveling waves that span across multiple brain areas. Brad Lega will discuss how theta phase organizes the firing of context-sensitive neurons, particularly time cells, to support recovery of context and content. Anli Liu will present evidence that ripple-locked spike patterns during naturalistic film viewing are reinstated during recall, supporting the idea that ripples contribute to successful retrieval by recovering scene-specific representations. Finally, Joel Voss will consider how neural oscillations are tightly linked to eye movements and other actions, raising the possibility of complex reciprocal interactions among memory-related neural dynamics and behavior. These findings in humans include many distinctive patterns that extend beyond what has been observed previously in animals. Together, these talks present human hippocampal oscillations as flexible organizing mechanisms that coordinate coding across space, time, meaning, and action, refining current models of episodic memory and motivating new approaches to study human memory function.

Presentations

Insights from direct human brain recordings on the neuronal coding of space and meaning

Josh Jacobs¹; ¹The University of Chicago

In this talk I will describe our work using direct human brain recordings to probe the novel aspects of the human brain for representing information and learning during spatial navigation and memory. First, I will show that place and grid cells in humans appear across a wider set of regions than in rodents, extending beyond the hippocampal-entorhinal circuit into parahippocampal, cingulate, and frontal cortex. Next, I will talk about how human hippocampal ripples coordinate place cell activity with non-spatial information, tying each location to the items and events encountered there, and how human place cells remap according to the meaning and task relevance of a location rather than its geometry alone. Finally, I will talk about the oscillations that organize this activity, including human theta rhythms, that run slower and in shorter bouts than in rodents, and low-frequency traveling waves that propagate across cortex and shift direction with behavioral demands. These findings demonstrate that the human brain shows novel patterns as well as recapitulates some of the results seen in rodents, which is important for constraining models of how memory and spatial navigation are implemented by the brain and for suggesting new paradigms for brain computer interfacing and neuromodulation.

Theta oscillations organize neural activity supporting episodic context recovery in the human MTL

Brad Lega¹; ¹The University of Texas Southwestern Medical Center

Temporally sensitive neurons provide a window into how brain oscillations organize contextual representations during memory. In spatial memory, place-cell sequences are structured within the theta cycle, suggesting that oscillatory phase may also organize human temporal context representations. We tested this idea using microelectrode recordings from the human hippocampus and entorhinal cortex while participants performed episodic memory tasks, including free recall and serial reconstruction. We analyzed single-neuron activity during encoding and retrieval of temporal context and classified neurons by firing dynamics as time cells, ramping cells, or non-time cells. Using the SPADE algorithm, we identified significant multi-neuron sequences and examined the relative spike timing of context-sensitive neurons around these sequence events. This approach was motivated by temporal context models proposing that retrieval depends on recovery of a prior contextual state, which then cues related memories according to temporal proximity. Temporal-cell populations showed systematic theta-phase organization during behavior. Time cells, but not ramping cells, exhibited preferred phase ordering within sequences, spiking earlier in the theta cycle relative to retrieval-related sequence events. These phase relationships predicted temporal clustering in recall, linking neuronal timing to memory organization. Phase organization of time cells was evident across the broad 2–10 Hz human theta range, consistent with the idea that multiple low-frequency hippocampal oscillations coordinate context-sensitive activity during memory search. Together, these findings support a model in which temporal context recovery is implemented through phase-organized interactions among time cells and larger neuronal sequences.

Ripple-locked spike patterns during film viewing are reinstated during recall

Anli Liu¹; ¹New York University Langone Health

Hippocampal ripples are brief, high-frequency oscillations thought to support memory by replaying behaviorally relevant spike sequences, but their role in human naturalistic behavior is poorly understood. We invited surgical epilepsy patients with invasive macro- and micro-wire recordings in medial temporal lobe (MTL) to watch a set of short films then later recall film content. During initial viewing, higher ripple rate and longer duration predict later retrieval of specific scenes. Using cross-validated linear classifiers on MTL spike trains, we show that ripple-locked population spiking reliably decodes film identity after a single exposure. During recall, ripple-associated spiking content more closely matches encoding-time activity for the recalled movie and scene compared to matched control periods. Together, these findings suggest that hippocampal ripples contribute to successful naturalistic memory retrieval in humans by reinstating scene-specific information.

Which came first, the behavior or the hippocampal oscillation?

Joel Voss¹; ¹The University of Chicago

There are many striking examples of oscillations in the human hippocampus that signal different qualities of episodic memories, such as theta and ripple oscillations that coordinate coding of space, time, and meaning. There is also a rich literature in rodents to indicate that hippocampal oscillations are tightly linked to simple behaviors. For example, theta phase-coding of future left versus right spatial trajectories are coordinated with left versus right stepping of the paws/feet, and theta generally is paced by oscillatory behaviors such as whisking, stepping, and iterative viewing. I will describe a similar phenomenon in humans, whereby hippocampal theta is tightly linked to saccadic eye movements. Saccades generate theta-band signals that are sensitive to memory qualities, suggesting that they may impose organization ON hippocampal memory-related processing. However, the timing of saccades is also predicted by the hippocampal theta rhythm, suggesting that they may BE organized BY memory-related processing. I argue that there is insufficient evidence to determine whether eye movements coordinate hippocampal theta, whether hippocampal theta coordinates eye movements, or whether both directions of influence co-exist. I will also discuss how this issue of circular causality concerns not just eye movements, but other behaviors that have been studied in relation to human hippocampal theta, particularly nasal respiration. Overall, I will argue that simple behaviors and memory-related neural activity patterns in the hippocampus are so tightly interwoven that we have not (yet) developed the tools to distinguish them, and will speculate as to various paths that could be taken to do so.