A novel method for quantifying similarities between oscillatory neural responses in wavelet time-frequency power profiles.

Quantifying similarities and differences between neural response patterns is an important step in understanding neural coding in sensory systems. It is difficult, however, to compare the degree of similarity among transient oscillatory responses. We developed a novel method of wavelet correlation analysis for quantifying similarity between transient oscillatory responses, and tested the method with olfactory cortical responses. In the anterior piriform cortex (aPC), the largest area of the primary olfactory cortex, odors induce inhibitory activities followed by transient oscillatory local field potentials (osci-LFPs). Qualitatively, the resulting time courses of osci-LFPs for identical odors were modestly different. We then compared several methods for quantifying the similarity between osci-LFPs for identical or different odors. Using fast Fourier transform band-pass filters, a conventional method demonstrated high correlations of the 0-2Hz components for both identical and different odors. None of the conventional methods tested demonstrated a clear correlation between osci-LFPs. However, wavelet correlation analysis resolved a stimulus dependency of 2-45Hz osci-LFPs in the aPC output layer, and produced experience-dependent high correlations in the input layer between some of the identical or different odors. These results suggest that redundancy in the neural representation of sensory information may change in the aPC. This wavelet correlation analysis may be useful for quantifying the similarities of transient oscillatory neural responses.