Data

Structured in the Neurodata Without Borders (NWB) format, this dataset spans 27 sessions involving 11 different animals. It yields data from 1,121 stable units, primarily putative excitatory but also including 126 putative inhibitory units. The subdataset we used includes detailed recordings from silicon probe electrodes implanted in the frontal cortices of male Long Evans rats, aged between 4-7 months. These recordings capture both local field potentials (LFPs) and spikes, providing a comprehensive view of neural dynamics over several days.

The LFP data serves as our primary measure.  LFPs are electric signals measured via an electrode within or directly on the surface the brain. They  measure the synaptic potentials across a group of neurons near the recording electrode. Typically, LFPs have high resolution in time that allows for the study of dynamic changes in the neurons. They also have characteristic frequency bands at which different levels of activity are noted:

• delta (0.5 - 4 Hz) : slow activity associated with deep sleep

• theta (4-8 Hz) : navigation, memory, and certain learning types

• alpha (8-12 Hz) : relaxed wakefulness and decreased attention

• beta (12-30 Hz) : active thinking, focus, and body movement

• gamma (31-200 Hz) : higher-order cognitive functions and conscious perception

Brain activity is tracked across various states and offers insights into the neural mechanisms that underpin different sleep cycles, focusing on transitions between wakefulness and sleep.

N-REM stage 1 involves the transition from alpha to theta waves. Alpha waves are produced with high amplitudes that become synchronized. N-REM stage 2 is comprised of higher amplitude theta waves with sleep spindles, bursts of higher frequency brain activity, and K-complexes (shown below). Last, in stage 3 of N-REM, we see high-amplitude delta waves.

REM sleep is comprised mostly of theta, beta, and gamma waves. These waves are low-amplitude and mixed frequency.

Last, awake brain waves are mostly noted in the beta and gamma waves, as well as the alpha waves. These waves are typically low-amplitude and higher frequency.

For our REM sleep classification project, we leveraged a specific portion of our extensive dataset, which comprises recordings from one day of a single male Long Evans rat. This targeted subset, accessed through MATLAB, concentrates on analyzing the transitions into REM sleep.

The dataset we used can be found at this link: https://dandiarchive.org/dandiset/000041/0.210812.1515/files?location=sub-Dino&page=1