mTRF-Toolbox is a MATLAB package for rapid estimation of forward encoding models (stimulus to neural response) or backward decoding models (neural response to stimulus), suitable for modelling neurophysiological data such as MEG, EEG, iEEG, sEEG, ECoG and EMG data.
Forward encoding models, also known as response functions or receptive fields, can be used to investigate information processing in neuronal populations with respect to temporal features (TRFs), or spectro- or spatio-temporal features (STRFs). STRFs can be subjected to conventional time-frequency and source analysis techniques used to analyse event related potentials. In addition, TRFs can be used to predict the dynamics of neural responses to unseen stimuli as a way to objectively measure stimulus encoding. Stimulus reconstruction can be performed using backward decoding models that project the multi-channel neural responses back to the dynamics of the stimulus. This is useful for decoding stimulus features from neural responses and can be used to build brain-computer interfaces and other real-time neural interface applications.
mTRF-Toolbox facilitates the use of natural continuous stimuli, allowing researchers to investigate how neural systems process dynamic environmental signals such as speech, music and motion, and to decode dynamic cognitive processes such as auditory attention and multisensory integration.
Crosse, Michael J., et al. “The Multivariate Temporal Response Function (MTRF) Toolbox: A MATLAB Toolbox for Relating Neural Signals to Continuous Stimuli.” Frontiers in Human Neuroscience, vol. 10, Frontiers Media SA, Nov. 2016, doi:10.3389/fnhum.2016.00604.