Decoding the spatio-temporal localization of behavioral events in populations of neurons

A method for decoding the spatio-temporal localization of behavioral events using the activity recorded from populations of neurons in-vivo is presented. The method is broken down into two parts: decoding temporal localization and decoding spatial localization of the events. As examples, the method is applied for (i) detecting the time of footfall contact during treadmill locomotion and (ii) classifying the location of passive cutaneous stimulation to the forepaw. The work presented was performed off-line but can easily be implemented in real-time to be used for neurorobotic control.