A color-based visualization technique for multielectrode spike trains.

Multi electrode recordings of neuronal activity provide an overwhelming amount of data that is often difficult to analyze and interpret. Although various methods exist for treating multielectrode datasets quantitatively, there is a particularly prominent lack of techniques that enable a quick visual exploration of such datasets. Here, by using Kohonen self-organizing maps, we propose a simple technique that allows for the representation of multiple spike trains through a sequence of color-coded population activity vectors. When multiple color sequences are grouped according to a certain criterion, e.g., by stimulation condition or recording time, one can inspect an entire dataset visually and extract quickly information about the identity, stimulus-locking and temporal distribution of multi-neuron activity patterns. Color sequences can be computed on various time scales revealing different aspects of the temporal dynamics and can emphasize high-order correlation patterns that are not detectable with pairwise techniques. Furthermore, this technique is useful for determining the stability of neuronal responses during a recording session. Due to its simplicity and reliance on perceptual grouping, the method is useful for both quick on-line visualization of incoming data and for more detailed post hoc analyses.

Hold Your Methods! How Multineuronal Firing Ensembles Can Be Studied Using Classical Spike-Train Analysis Techniques.

Responses of neuronal populations play an important role in the encoding of stimulus related information. However, the inherent multidimensionality required to describe population activity has imposed significant challenges and has limited the applicability of classical spike train analysis techniques. Here, we show that these limitations can be overcome. We first quantify the collective activity of neurons as multidimensional vectors (patterns). Then we characterize the behavior of these patterns by applying classical spike train analysis techniques: peri-stimulus time histograms, tuning curves and auto- and cross-correlation histograms. We find that patterns can exhibit a broad spectrum of properties, some resembling and others substantially differing from those of their component neurons. We show that in some cases pattern behavior cannot be intuitively inferred from the activity of component neurons. Importantly, silent neurons play a critical role in shaping pattern expression. By correlating pattern timing with local-field potentials, we show that the method can reveal fine temporal coordination of cortical circuits at the mesoscale. Because of its simplicity and reliance on well understood classical analysis methods the proposed approach is valuable for the study of neuronal population dynamics.

Timescales of multineuronal activity patterns reflect temporal structure of visual stimuli.

The investigation of distributed coding across multiple neurons in the cortex remains to this date a challenge. Our current understanding of collective encoding of information and the relevant timescales is still limited. Most results are restricted to disparate timescales, focused on either very fast, e.g., spike-synchrony, or slow timescales, e.g., firing rate. Here, we investigated systematically multineuronal activity patterns evolving on different timescales, spanning the whole range from spike-synchrony to mean firing rate. Using multi-electrode recordings from cat visual cortex, we show that cortical responses can be described as trajectories in a high-dimensional pattern space. Patterns evolve on a continuum of coexisting timescales that strongly relate to the temporal properties of stimuli. Timescales consistent with the time constants of neuronal membranes and fast synaptic transmission (5-20 ms) play a particularly salient role in encoding a large amount of stimulus-related information. Thus, to faithfully encode the properties of visual stimuli the brain engages multiple neurons into activity patterns evolving on multiple timescales.

The oscillation score: an efficient method for estimating oscillation strength in neuronal activity.

We present a method that estimates the strength of neuronal oscillations at the cellular level, relying on autocorrelation histograms computed on spike trains. The method delivers a number, termed oscillation score, that estimates the degree to which a neuron is oscillating in a given frequency band. Moreover, it can also reliably identify the oscillation frequency and strength in the given band, independently of the oscillation in other frequency bands, and thus it can handle superimposed oscillations on multiple scales (theta, alpha, beta, gamma, etc.). The method is relatively simple and fast. It can cope with a low number of spikes, converging exponentially fast with the number of spikes, to a stable estimation of the oscillation strength. It thus lends itself to the analysis of spike-sorted single-unit activity from electrophysiological recordings. We show that the method performs well on experimental data recorded from cat visual cortex and also compares favorably to other methods. In addition, we provide a measure, termed confidence score, that determines the stability of the oscillation score estimate over trials.