Interspike interval analysis and spikelets in presubicular head-direction cells

Stefano Coletta, Roxana Zeraati, Khaled Nasr, Patricia Preston‐Ferrer, Andrea BurgalossiJournal of Neurophysiology 2018 · DOI

TL;DR

This study reveals that head-direction (HD) cells in the presubiculum, crucial for navigation, can be categorized into two distinct types based on their firing patterns: "bursty" and "nonbursty." The researchers also identified "spikelets," small electrical events, in a subset of these cells, suggesting these features might enhance the precision and reliability of spatial information coding.

What Problem Does This Paper Address?

Navigating effectively requires an internal sense of direction, a function largely attributed to head-direction (HD) cells. While it's known that these cells fire more when an animal's head is oriented in a specific direction, and that their firing rate encodes this information, the precise mechanisms by which this directional signal is encoded and transmitted remain an active area of research. Specifically, the role of fine temporal features within the spike trains, beyond just the firing rate, is not fully understood. This paper investigates whether variations in spike timing, such as "bursts" (rapid sequences of spikes) and "spikelets" (small, spike-like events), contribute to the coding and robustness of the HD signal in the presubiculum, a key brain region for spatial navigation.

Methods

The researchers employed juxtacellular recordings in head-fixed rats undergoing passive rotation, a standard technique for studying HD cells in a controlled environment. They also conducted recordings in freely moving rats to observe HD cell activity during natural behavior. A key methodological contribution is the use of interspike interval analysis and principal component analysis (PCA) to classify HD neurons into "bursty" and "nonbursty" groups based on their tendency to fire spikes with very short intervals. Additionally, they identified and characterized "spikelets" by analyzing their amplitude and kinetics, comparing them to actual spikes. To confirm the origin of spikelets, they used a brief current "buzz" to achieve partial intracellular access and also performed whole-cell recordings in awake rats.

Datasets

The study utilized electrophysiological recordings from 201 presubicular HD cells in passively rotated rats and 18 cells from freely moving rats. These datasets were collected as part of the study and are not publicly released benchmarks.

Key Insights and Ideas

  • HD Cells Aren't Monolithic: The discovery that presubicular HD cells can be reliably classified into "bursty" and "nonbursty" subtypes based on their firing patterns challenges the idea of a uniform HD cell population.
  • Bursty Cells are "Better" Navigators: Bursty neurons, characterized by distinct spike waveforms and a propensity for short interspike intervals, exhibited stronger directional tuning and higher signal-to-noise ratios, suggesting they might play a more critical role in precise HD coding.
  • Spikelets: A Hidden Signal: The identification of spikelets, small electrical events, in a subset of HD cells, and their highly correlated HD tuning with the main spikes, suggests they carry meaningful spatial information.
  • Juxtacellular Recordings Unlock New Insights: The ability to reliably detect spikelets using juxtacellular recordings (previously challenging due to their small amplitude) opens new avenues for studying their role in freely behaving animals.
  • Temporal Coding Matters: The findings strongly suggest that the temporal structure of neuronal firing, not just the rate, is crucial for encoding complex information like head direction.

Key Results

  • Two Cell Classes: 201 presubicular HD cells were classified into two groups: 64 "bursty" and 137 "nonbursty."
  • Bursty Cell Superiority: Bursty neurons showed significantly higher HD indexes (0.87 vs. 0.80), narrower tuning curve widths (70.6° vs. 79.8°), and larger signal-to-noise ratios (230.8 vs. 93.8) compared to nonbursty cells.
  • Spikelet Occurrence: Spikelets were observed in approximately 7% of recorded HD cells (15 out of 201).
  • Spikelet Tuning: The HD tuning of spikelets was virtually identical to that of the main spikes in the same neuron.
  • Spikelets Enhance Firing Rate: Including spikelets in the firing rate calculation significantly increased the measured rate (31.8 Hz vs. 17.2 Hz).
  • Bursty and Spikelet Correlation: Neurons exhibiting spikelets also tended to have a higher burst index, suggesting a potential link between these phenomena.
  • Natural Behavior Confirmation: The distinction between bursty and nonbursty cells, and the presence of spikelets, were also observed in freely moving animals during natural exploration.

Limitations and Future Work

The authors acknowledge that while their evidence strongly suggests spikelets originate from the same neuron, they cannot definitively rule out the possibility of them being spikes from a second, co-tuned neuron, especially given the limitations of extracellular recordings. Future work could involve functional imaging to resolve the precise topographical organization of HD cells and their connections. Further experiments are also needed to elucidate the exact mechanisms underlying spikelet generation (e.g., failed axonal spikes vs. electrotonic coupling) and to fully understand their role in synaptic plasticity and information transmission to downstream targets like the medial entorhinal cortex. The study also suggests exploring whether other HD cell properties, like stability and flexibility in response to environmental changes, map onto the identified bursty/nonbursty classes.

Summary updated

Abstract

Head-direction (HD) neurons are thought to provide the mammalian brain with an internal sense of direction. These cells, which selectively increase their firing when the animal's head points in a specific direction, use the spike rate to encode HD with a high signal-to-noise ratio. In the present work, we analyzed spike train features of presubicular HD cells recorded juxtacellularly in passively rotated rats. We found that HD neurons could be classified into two groups on the basis of their propensity to fire spikes at short interspike intervals. "Bursty" neurons displayed distinct spike waveforms and were weakly but significantly more modulated by HD compared with "nonbursty" cells. In a subset of HD neurons, we observed the occurrence of spikelets, small-amplitude "spike-like" events, whose HD tuning was highly correlated to that of the co-recorded juxtacellular spikes. Bursty and nonbursty HD cells, as well as spikelets, were also observed in freely moving animals during natural behavior. We speculate that spike bursts and spikelets might contribute to presubicular HD coding by enhancing its accuracy and transmission reliability to downstream targets. NEW & NOTEWORTHY We provide evidence that presubicular head-direction (HD) cells can be classified into two classes (bursty and nonbursty) on the basis of their propensity to fire spikes at short interspike intervals. Bursty cells displayed distinct electrophysiological properties and stronger directional tuning compared with nonbursty neurons. We also provide evidence for the occurrence of spikelets in a subset of HD cells. These electrophysiological features (spike bursts and spikelets) might contribute to the precision and robustness of the presubicular HD code.
BibTeX
@article{coletta2018interspike,
  author = {Stefano Coletta and Roxana Zeraati and Khaled Nasr and Patricia Preston‐Ferrer and Andrea Burgalossi},
  title = {{Interspike interval analysis and spikelets in presubicular head-direction cells}},
  journal = {Journal of Neurophysiology},
  year = {2018},
  doi = {10.1152/jn.00019.2018},
}
Figure