Time-Frequency Analysis of Error Related Activity in Anterior Cingulate Cortex

摘要

Error related activity in local field potentials recorded in dorsal anterior cingulate cortex (dACC) of non-human primates is analyzed using ensemble empirical mode decomposition (EEMD). Neuronal activity in this region was recorded in a male rhesus monkey trained in the problem solving task (PS). It was hypothesized that activity in this region could reflect a modification of control just before execution errors, hence predicting the error about to be committed. The features obtained from time-frequency energy distribution, intrinsic mode functions (IMF) and higher order statistics calculated over IMFs were used to distinguish error related activity. The analysis has revealed the significant increase in energy of IMF4-IMF7, resulting in significant total energy increase immediately before and after execution errors. Higher order spectra analysis determined no difference in symmetry and shape of sample distribution when compared to control segments from successful trials. However, increased sample variance during errors combined with increase in energy indicates the unstable neural activation in this cortical region during execution errors resulting in less control over the current processing.