Sparse high-dimensional decomposition of non-primary auditory cortical receptive fields.

Area of Science:

• Neuroscience
• Auditory System
• Computational Neuroscience

Background:

• Characterizing neuronal responses to natural stimuli is crucial in sensory neuroscience.
• Spectrotemporal receptive fields (STRFs) summarize auditory cortical neuron responses but are complex in non-primary areas.
• Understanding auditory pathway transformations requires deciphering complex non-primary STRFs.

Purpose of the Study:

• To investigate the relationship between ferret primary auditory cortex (A1) and secondary dorsal posterior ectosylvian gyrus (PEG).
• To propose and apply a novel method (Cortical Receptive Fields - CortRF) for estimating receptive fields in PEG.
• To understand how acoustic stimulus representations are transformed along the auditory pathway.

Main Methods:

• Estimated receptive fields in PEG using a high-dimensional computational model of primary-cortical stimulus representations (CortRF).
• Applied CortRF analysis to neuronal responses in PEG and A1 using speech and temporally orthogonal ripple combination (TORC) stimuli.
• Greedily identified salient primary-cortical features modulating spiking responses in PEG.

Main Results:

• CortRFs of PEG neurons captured selectivity to more complex spectrotemporal features than A1 neurons.
• CortRF models were more predictive of PEG responses to speech compared to A1.
• Incorporating primary-cortical representations improved the prediction of PEG single-unit responses to natural sounds.

Conclusions:

• Secondary-cortical stimulus representations in PEG are computed as sparse combinations of primary-cortical features.
• This hierarchical decomposition facilitates the encoding of natural stimuli in the auditory pathway.
• The findings explicitly confirm the presumed hierarchical organization of the auditory cortex.