PAIMD: A Novel Data-Driven Decomposition Method for Separating Neural Signal Into Periodic and Aperiodic Components

PAIMD: A Novel Data-Driven Decomposition Method for Separating Neural Signal Into Periodic and Aperiodic Components

Blog Article

Recording of brain activity comprises both periodic and aperiodic components.Recent studies postulate distinct physiological interpretations for aperiodic activity.Ignoring the aperiodic (fractal) component when analyzing the brain signal could result in incorrect conclusions.We proposed a new data-driven solution to the signal decomposition problem, called periodic-aperiodic intrinsic mode decomposition (PAIMD), for 12n/1200 wella separating the brain signal into the oscillations (periodic activity) and underlying fractal background (aperiodic activity).

The PAIMD does not need any specification for the frequency limits of oscillations.Furthermore, a new technique was introduced for calculating the peaks in the spectrum of signals which feed to the PAIMD as the initial condition.We tested the PAIMD on simulated data for signal decomposition when the fractal activity was present.In addition, the application of PAIMD on human EEG data during the delay period of the working memory task revealed alternation in fractal features (exponent and offset) for individuals with lateral PFC damage and task-related alternation in alpha-beta oscillation power gtech brush bar (relation and identity features of visual stimuli).