Human sensorimotor resting state beta events and 1/f response show good test-retest reliability

Neurological conditions affecting the sensorimotor system have a profound impact on individuals physical independence and are associated with a considerable socioeconomic burden. Reliable functional biomarkers allowing early diagnosis of these conditions or targeting treatment and rehabilitation can reduce this burden. Magnetoencephalography (MEG) can non-invasively measure the brains salient rhythmic patterns such as the somatomotor ( rolandic) rhythm. This rhythm shows intermittent high amplitude events in the beta (14-30 Hz) frequency range which predict behavior across tasks and species and are altered by neurological diseases affecting the sensorimotor system. Thus, the sensorimotor resting beta phenotype is a promising candidate biomarker of sensorimotor function. A prerequisite for use as a biomarker is that it can be quantified reliably across different measurement sessions. Here, using MEG, we assessed the test-retest stability of spontaneously occurring sensorimotor power spectral characteristics, including both aperiodic (1/f) as well as beta band fluctuations ( beta events) in a cohort of 50 healthy human controls. Test-retest reliability across two separate measurement sessions was assessed using the intraclass correlation coefficient (ICC). Beta events were determined using a thresholding-based approach on a narrow-band filtered amplitude envelope obtained using Morlet wavelet decomposition across a range of parameters (recording length, amplitude threshold and filtering bandwidth). We find that both aperiodic power spectral features as well as several beta event characteristics show good to excellent testretest stability. Especially aperiodic component power spectral features (ICC 0.77-0.88), but also measures of beta event amplitude (ICC 0.74-0.82) were found to be very stable, while measures of individual beta event duration were less reliable, especially for the left hemisphere (ICC right [~]0.7, left [~]0.55). Recordings of 2-3 minutes were sufficient to obtain stable results for most parameters. Important for potential clinical applications, automatization of beta event extraction was successful in 86 % of cases. Beta event rate and duration measures were more sensitive to analysis parameters than the measures of event amplitude. The results suggest the sensorimotor beta phenotype is a stable feature of an individuals resting brain activity even for short, 2-3 minute recordings which can be easily measured in patient populations, facilitating its use as a potential clinical biomarker.