Proximal CA1 20–40 Hz power dynamics reflect trial-specific information processing supporting nonspatial sequence memory

Sandra Gattas, Gabriel A. Elias, John Janecek, Michael A. Yassa, Norbert J. FortineLife 2022 · DOI

TL;DR

This study reveals that specific brainwave patterns (20-40 Hz oscillations) in the hippocampus, particularly in its proximal region, are crucial for remembering the order of nonspatial events. These brainwave dynamics predict memory performance and appear to act as a "gate," protecting important information during decision-making.

What Problem Does This Paper Address?

The hippocampus is well-known for its role in spatial memory, but its involvement in processing temporal information, especially the order of events, is less understood. While brain oscillations (rhythmic electrical activity) are implicated in memory, how specific oscillatory patterns in the hippocampus contribute to remembering the sequence of nonspatial information remains unclear. This paper investigates how hippocampal oscillations, specifically in the CA1 region, are involved in processing temporal context for nonspatial memories and how these oscillations vary across different parts of CA1.

Methods

The researchers recorded local field potential (LFP) activity from the CA1 region of the hippocampus in rats as they performed a nonspatial odor sequence memory task. This task required rats to remember the order of five different odors. They used tetrodes to record from multiple locations along the proximodistal axis of CA1, allowing them to examine how brain activity differed across this anatomical gradient. Spectral analysis, including time-frequency analysis, was used to examine oscillatory power in different frequency bands. Statistical analyses, including linear mixed-effects models (LMEM ANOVA) and cluster-based permutation testing (CBPT), were employed to identify significant differences in brainwave activity related to task performance, behavioral states (odor processing vs. running), and anatomical location. A random forest classifier was also used to predict trial accuracy based on 20-40 Hz power.

Datasets

The study used data collected from five male Long-Evans rats. The dataset consists of LFP recordings from these animals during a well-trained session and two novel sequence learning sessions. No new benchmark datasets were introduced; the data was collected specifically for this study.

Key Insights and Ideas

  • 20-40 Hz oscillations are key for nonspatial sequence memory: The study strongly suggests that these specific brainwave frequencies in CA1 are not just correlated with memory but are actively involved in processing the temporal order of nonspatial events.
  • Anatomical specialization within CA1: Different parts of CA1 show distinct oscillatory patterns. 20-40 Hz power is higher in proximal CA1 during odor processing, while theta power is higher in distal CA1. This suggests a functional segregation within the CA1 region.
  • "Gating" mechanism: The sustained 20-40 Hz power observed during "hold-response" trials (where animals had to maintain a decision) points to a role in protecting and maintaining task-relevant information, acting like a gate to prevent interference.
  • Performance prediction: The strength of 20-40 Hz power in proximal CA1, particularly during the sustained "hold" period, directly predicts the accuracy of the animal's memory judgments.
  • Learning-dependent changes: 20-40 Hz power increases with learning and performance level, suggesting it reflects the consolidation of sequence knowledge.

Key Results

  • Odor sequence processing epochs showed distinct spectral profiles and proximodistal CA1 gradients of theta and 20-40 Hz power compared to track running epochs.
  • 20-40 Hz power increased towards proximal CA1 during odor processing, while theta power increased towards distal CA1.
  • 20-40 Hz power was positively correlated with performance level across sessions, with significantly higher power in well-trained sessions compared to novel ones.
  • In "hold-response" trials, 20-40 Hz power reached a steady-state plateau, which was predictive of trial accuracy. Incorrect judgments in these trials were associated with higher 20-40 Hz power, suggesting over-suppression or interference.
  • A random forest classifier could predict trial accuracy using 20-40 Hz power from proximal CA1 during a late trial epoch, achieving an AUC of 0.588.

Limitations and Future Work

  • Response type artifact: The study acknowledges that the distinction between "hold" and "withdraw" responses, while crucial for the task, might have introduced some differences in power dynamics beyond just the decision itself, potentially due to variations in cue utilization or trial duration. Future work could explore this further with different experimental designs.
  • Uneven trial distribution: The task design required a lower proportion of "out of sequence" trials, leading to an uneven distribution of data that could influence analyses. The authors controlled for this using statistical methods.
  • Origin of oscillations: The exact neural circuits responsible for generating these 20-40 Hz oscillations in CA1 remain unknown. Future research should investigate the contributions of upstream structures like the entorhinal cortex and medial septum.
  • Multiregional integration: While the study highlights CA1's role, future work should further explore how these beta oscillations coordinate activity across different brain regions to support complex temporal memory.

Summary updated

Abstract

The hippocampus is known to play a critical role in processing information about temporal context. However, it remains unclear how hippocampal oscillations are involved, and how their functional organization is influenced by connectivity gradients. We examined local field potential activity in CA1 as rats performed a nonspatial odor sequence memory task. We found that odor sequence processing epochs were characterized by distinct spectral profiles and proximodistal CA1 gradients of theta and 20-40 Hz power than track running epochs. We also discovered that 20-40 Hz power was predictive of sequence memory performance, particularly in proximal CA1 and during the plateau of high power observed in trials in which animals had to maintain their decision until instructed to respond. Altogether, these results provide evidence that dynamics of 20-40 Hz power along the CA1 axis are linked to trial-specific processing of nonspatial information critical to order judgments and are consistent with a role for 20-40 Hz power in gating information processing.
BibTeX
@inproceedings{gattas2022proximal,
  author = {Sandra Gattas and Gabriel A. Elias and John Janecek and Michael A. Yassa and Norbert J. Fortin},
  title = {{Proximal CA1 20–40 Hz power dynamics reflect trial-specific information processing supporting nonspatial sequence memory}},
  booktitle = {eLife},
  year = {2022},
  doi = {10.7554/elife.55528},
}
Figure