Oscillations Go the Distance: Low-Frequency Human Hippocampal Oscillations Code Spatial Distance in the Absence of Sensory Cues during Teleportation

Abstract

Low-frequency (delta/theta band) hippocampal neural oscillations play prominent roles in computational models of spatial navigation, but their exact function remains unknown. Some theories propose they are primarily generated in response to sensorimotor processing, while others suggest a role in memory-related processing. We directly recorded hippocampal EEG activity in patients undergoing seizure monitoring while they explored a virtual environment containing teleporters. Critically, this manipulation allowed patients to experience movement through space in the absence of visual and self-motion cues. The prevalence and duration of low-frequency hippocampal oscillations were unchanged by this manipulation, indicating that sensorimotor processing was not required to elicit them during navigation. Furthermore, the frequency-wise pattern of oscillation prevalence during teleportation contained spatial information capable of classifying the distance teleported. These results demonstrate that movement-related sensory information is not required to drive spatially informative low-frequency hippocampal oscillations during navigation and suggest a specific function in memory-related spatial updating.

Figure 1

Environment layout and example trial structure. Top left, top-down view of the plus maze, with target stores circled in red and teleporters circled in blue. This view is for schematic purposes only and was not shown to patients. Top right, views from one example trial. Each trial starts in the central plaza (1) where patients first find the landmark associated with the target store, which is visible in the distance, but whose identity is obscured by fog. Patients proceed down the arm of the maze (2) until reaching the target store (3). They then proceed to the nearest teleporter (4), at which point the display fades to black (5) and they teleport to the central plaza (6) to begin the next trial. Example trial shown is a short teleportation distance. Bottom, schematic of the relative timing of navigation and teleportation epochs. Teleportation epochs were time-locked to the moments of teleporter entry and exit, and were either 1830 ms or 2830 ms in duration. Navigation epochs started after the patient entered the correct arm of the maze and were matched in duration to the upcoming teleportation epochs.

Figure 2

Oscillatory activity by time point and condition. A) Example recordings during navigation and teleportation in three electrodes. Raw traces (left) show iEEG activity over time for the three teleportation epochs (1, 2, 3; dashed blue lines represent teleporter entry and exit) and their matched navigation epochs (1, 2, 3; dashed blue lines represent artificially imposed boundaries). Center, mean Delta-Theta P_Episode for each epoch of the example trial. Right, mean Delta-Theta P_Episode averaged across all trials for that electrode. B) Mean P_Episode values for each epoch for each frequency band. Each line represents one electrode. Overall, low-frequency P_Episode does not differ across epochs. C) Difference in mean P_Episode between active navigation and teleportation for each pair of matched epochs (1, 2, 3, as in A). A significant proportion of electrodes show reduced Delta-Theta and Alpha P_Episode post-teleportation (Epoch 3), likely because navigation speed is much slower during this time. Data shown as mean ± SEM. See also Figures S1 and S2, Tables S1 and S2.

Figure 3

Oscillatory episode duration for navigation and teleportation. Each line is an electrode and indicates the mean (± SEM) duration of oscillatory episodes that cross either the time point of teleporter entry or an artificially imposed boundary during navigation (see Figure 2A). Duration refers to the length of time the oscillation persisted after entering the teleporter or crossing the artificial boundary. No electrodes exhibited a significant difference in Delta-Theta oscillatory episode duration, indicating that entering the teleporter did not disrupt ongoing oscillations. See also Figure S3.

Figure 4

Successful classification of distance teleported using low-frequency P_Episode. Left, raw traces from two example electrodes showing individual short-distance (top) and long-distance (bottom) teleportation events. Colored bars below each trace indicate periods of significant oscillatory episodes for each frequency. Color scale at right indicates frequencies in Hz. In both electrodes, there is greater low-frequency oscillatory activity during the long-distance teleportation event. Center, for the same electrodes, mean P_Episode at each frequency during short-distance and long-distance teleportation trials. Right, for the same electrodes, classification results for each trial across iterations. Perfect classification over all iterations would be characterized by all purple bars at 0% and all green bars at 100%. Text at right indicates the number of trials that achieved correct classification for >50% of iterations. For display purposes, we show the classification of individual trials across iterations; however, statistical analyses were performed on the mean classification across trials for each iteration.