Document Type : Full Research Paper
Authors
1 MSc. Student, Biomedical Engineering Department, Electrical Engineering Faculty, Iran University of Science and Technology, Tehran, Iran
2 Ph.D Student, Biomedical Engineering Department, Electrical Engineering Faculty, Iran University of Science and Technology, Tehran, Iran
3 Assistant Professor, Biomedical Engineering Department, Electrical Engineering Faculty, Iran University of Science and Technology, Tehran, Iran
4 Professor, Biomedical Engineering Department, Electrical Engineering Faculty, Iran University of Science and Technology, Tehran, Iran
Abstract
Neural adaptation is a brain ability which reduces the neural activities in response to a repeated stimulus. In this study, we examined the effect of adaptation on neural decoding. For this purpose, pure tones with different frequency-amplitude combinations were presented randomly in two sequences (usual and adaptive). During the task, local field potential (LFP) signals were recorded from the primary auditory cortex of fifteen anesthetized rats. In the usual sequence, the stimuli were presented randomly with 50 ms duration and 300 ms interstimulus interval (ISI). Each combination was presented about 25 times. In the adaptive sequence, same as the usual one, stimuli were presented with this difference that one specific frequency (adapter) with the probability of 80% was presented frequently in this sequence. Comparison between decoding accuracy of two sequences allows us to study the effect of adaptation to a specific frequency on neural decoding. First, considering the power spectrum feature in six frequency bands and using LDA (linear discrimination analysis) classifier, the average decoding accuracy of all frequency-pairs were calculated in the usual sequence. Subsequently, the decoding accuracy of frequency-pairs in the adaptive sequence was calculated and compared with the usual sequence. Results show a significant decoding accuracy between different frequency-pairs in beta, gamma, and high-gamma bands (>12 Hz) of local field potential with an accuracy of about 80%. Moreover, we found that adaptation to one frequency of sound decreases the decoding accuracy of neighbor frequencies. This signature was observed in high-frequency gamma and high-gamma activities (30-120 Hz) of LFPs.
Keywords