Extracting information in spike time patterns with wavelets and information theory

Abstract

Lopes-dos-Santos V, Panzeri S, Kayser C, Diamond ME, Quian Quiroga R. Extracting information in spike time patterns with wavelets and information theory. J Neurophysiol 113: 1015-1033, 2015. First published November 12, 2014 doi:10.1152/jn.00380.2014.-We present a new method to assess the information carried by temporal patterns in spike trains. The method first performs a wavelet decomposition of the spike trains, then uses Shannon information to select a subset of coefficients carrying information, and finally assesses timing information in terms of decoding performance: the ability to identify the presented stimuli from spike train patterns. We show that the method allows: 1) a robust assessment of the information carried by spike time patterns even when this is distributed across multiple time scales and time points; 2) an effective denoising of the raster plots that improves the estimate of stimulus tuning of spike trains; and 3) an assessment of the information carried by temporally coordinated spikes across neurons. Using simulated data, we demonstrate that the Wavelet-Information (WI) method performs better and is more robust to spike time-jitter, background noise, and sample size than wellestablished approaches, such as principal component analysis, direct estimates of information from digitized spike trains, or a metric-based method. Furthermore, when applied to real spike trains from monkey auditory cortex and from rat barrel cortex, the WI method allows extracting larger amounts of spike timing information. Importantly, the fact that the WI method incorporates multiple time scales makes it robust to the choice of partly arbitrary parameters such as temporal resolution, response window length, number of response features considered, and the number of available trials. These results highlight the potential of the proposed method for accurate and objective assessments of how spike timing encodes information. neuronal coding; temporal coding; wavelets; information theory; decoding THE IMPORTANCE OF PRECISE spike timing in carrying meaningful information has attracted much attention (Quian Quiroga and Panzeri 2009;Rieke et al. 1999). Does the temporal structure of spike trains provide information beyond the total spike count, or does it merely reflect noise? According to the "rate coding" view, neurons represent stimuli solely by the rate of firing within an encoding time window (Adrian and Zotterman 1926;Shadlen and Newsome 1994). In contrast, according to the "temporal coding" view, the time structure of the responses conveys additional information not provided by the total spike count (de Ruyter van Steveninck et al. 1997;Optican and Richmond 1987;Richmond and Optican 1987;Victor and Purpura 1996).Experimental evidence accumulated over the last 3 decades has suggested that precise spike patterns, on the scale of milliseconds, do indeed convey information not available in rate codes (Arabzadeh et al. 2006;de Ruyter van Steveninck et al. 1997;Di Lorenzo et al. 2009;Eckhorn and Popel 1975;Foffani et al. ...

On Information Metrics for Spatial Coding

“…In other words, if two variables are related, it might be possible to use one of them to decode the other. The information contained in the confusion matrix of a decoder provides a lower bound to the information between the two variables (Quiroga and Panzeri, 2009), allowing the use of decoding performance as an empirical estimate of the real (spatial) information of the cell (Robertson et al, 1999;Jensen and Lisman, 2000;Huxter et al, 2008;Lopes-dos-Santos et al, 2015). In this work, the performance of a Bayesian decoder was assumed to represent the gold standard of the true spatial information content of a cell.…”

“…Using information-theoretic metrics has been a frequent strategy to investigate neuronal encoding in different brain areas. For instance, the MI has been previously applied to a variety of experimental paradigms aimed at identifying the neuronal correlates of sensory and motor functions as well as of high-level processes such as working memory (Panzeri et al, 2001;Petersen et al, 2001;Belitski et al, 2008Belitski et al, , 2010Montemurro et al, 2008;Kayser et al, 2009;Pava˜o et al, 2014;Lopes-dos-Santos et al, 2015;Rossi-Pool et al, 2016;Vergara et al, 2016). Moreover, the MI has also been used to study diverse coding strategies, such as the amount of information carried by different types of signals (e.g., LFP oscillations and spikes) and/or by a combination of their features (Belitski et al, 2008(Belitski et al, , 2010Montemurro et al, 2008;Kayser et al, 2009;Pava˜o et al, 2014).…”

Neuronal Assemblies Evidence Distributed Interactions within a Tactile Discrimination Task in Rats

“…that NA units carry more task-related information than non-NA units) support that this approach can reveal physiologically relevant neural correlations. In this way, methods that incorporate information theory (Lopes-dos Santos et al, 2015 ) or statistical frameworks (Russo and Durstewitz, 2017 ) may enhance the assessment of distributed cortical interactions. Nevertheless, similarly to previous works (Peyrache et al, 2010 ; Lopes-dos Santos et al, 2011 , 2013 ; Almeida-Filho et al, 2014 ; Gulati et al, 2014 ; Bower et al, 2015 ), the PCA+ICA method for characterizing NAs revealed the formation of task-related groups of neurons that are more informative about tactile discrimination than random groups of neurons, which is suggestive of functional neural interactions.…”

On information metrics for spatial coding

“…In other words, if two variables are related, it might be possible to use one of them to decode the other. The information contained in the confusion matrix of a decoder provides a lower bound to the information between the two variables (Quiroga and Panzeri, 2009), allowing the use of decoding performance as an empirical estimate of the real (spatial) information of the cell (Robertson et al, 1999;Jensen and Lisman, 2000;Huxter et al, 2008;Lopes-dos-Santos et al, 2015). Under this framework, our simulations show that the mutual information (MI) better correlates with spatial decoding performance than I sec and I spike .…”