The code provided is part of a computational neuroscience model designed to analyze neural data and extract key metrics related to neuronal activity. The focus is on quantifying aspects of single-unit activity (SUA) and multi-unit activity (MUA) within neural populations. Below are the biological concepts and processes that the code attempts to model or analyze:
The calculation of firing rates (dsCalcFRmulti) is a fundamental measure of neuronal activity reflecting the frequency at which neurons generate action potentials (or spikes). Firing rates are indicative of how neurons encode information through their spiking activity. The code calculates firing rates over different time bins (e.g., 1-second and 100-millisecond bins) to capture both long-term patterns and transient spiking behavior.
The code optionally calculates the power spectral density (dsCalcPower), which is a measure of signal power distributed across different frequency bands. This can help identify oscillatory patterns in neural data, such as gamma or theta rhythms, which have been associated with various cognitive and sensory processes.
Interspike intervals represent the time between consecutive spikes. The analysis of ISIs (dsCalcISI) can reveal information about neuronal firing patterns, such as bursting behavior or regular spiking. Bursting, identified via Gaussian mixture models in the code, often correlates with certain neuronal states or functions, such as information transmission in certain brain regions.
The auto-correlation function analysis (dsCalcACF) provides insights into rhythmic spiking behavior and temporal patterns of spikes. It can identify periodicities and synchronization in neuronal firing, which may relate to coherent activity across neural circuits.
Synchronization between SUA and MUA is evaluated to assess the degree of coherence in spiking between individual neurons and the larger population (multi-unit activity). Differences in firing rates and power spectral properties between SUA and MUA might indicate variations in neuronal excitability or input integration across the population.
The code attempts to detect burst firing, which is characterized by a rapid succession of spikes. Burst firing can play a crucial role in synaptic plasticity and signal amplification. Bursts are detected based on distribution characteristics of ISIs, using AIC from Gaussian mixture models to differentiate between unimodal (non-bursting) and bimodal (bursting) ISI distributions.
The code evaluates periods of silence in neuronal firing by analyzing long ISIs beyond the mean and standard deviation of the ISI distribution. These silent periods can indicate neuronal inactivation or hyperpolarization, which may be crucial for understanding inhibitory control within neural circuits.
The median voltage of the membrane potential (median(data.([thisPop '_V']))) captures the baseline electrical state of neurons. This helps in understanding the resting state and how neurons respond to inputs.
Overall, the code is designed to assess key aspects of neural activity — such as firing rates, oscillations, bursting, synchronization, and silent periods — which are all critical for understanding how neuronal circuits function and how they may underlie behavior and cognitive processes. These metrics provide valuable insights into the functional properties of neural populations, contributing to our understanding of brain dynamics in health and disease.