BRAIN RESEARCH ELSEVIER Brain Research 734 (1996) 252-260 Research report Neuromagnetic study of movement-related changes in rhythmic brain activity Bernd Feige a,,, Rumyana Kristeva-Feige b, Simone Rossi c, Vittorio Pizzella d, Paolo-Maria Rossini e,f a Department of Psychiatry, University ofFreiburg, Hauptstrafle 5, 79104 Freiburg, Germany b Neurological Unicersity Clinic, University ofFreiburg, Freiburg, Germany c lstituto di Clinica delle Malattie Nervose e Mentali, Universitd di Siena, Siena, Italy d Istituto di Elettronica dello Stato Solido, CNR, Roma, Italy e DiHsione Neurologia, Ospedale "Fatebenefratelli', lsola Tiberina, Roma, Italy f I.R.C.C.S., 'Santa Lucia', Via Ardeatina, Roma, ltaly Accepted 15 May 1996 Abstract Neuromagnetic fields from the left cerebral hemisphere of five healthy, right-handed subjects were investigated in a typical Bereitschaftspotential paradigm consisting of self-paced voluntary movement of the right index finger. To assess movement-related spectral changes of the spontaneous magnetoencephalogram, latency- dependent short-time spectra were obtained by Fourier analysis for each single trial. The number of trials in which the spectral estimate for a certain frequency and latency deviated from reference values was then transformed into a probabilistic relative power measure. A spectral power depression around 20 Hz was observed starting about 2.5 s before movement onset, followed by elevated power in the 20-35 Hz range starting about 500 ms after movement onset. Generally, the power increase differed from the prior depression in both spectrum and topography, suggesting different generating processes rather than just a 'rebound' effect of the idling rhythm generator. The time course and topography of spectral power changes are discussed in relation to the corresponding properties of the movement-related neuromagnetic fields (readiness field, motor field, and movement-evoked field I). Keywords: Movement; Human; Magnetoencephalography;Spectral power; Spontaneous activity; Motor cortex; Somatosensorycortex; Cortical rhythm; Event-related spectral change; Movement-related magnetic field 1. Introduction Most electrical macropotential and neuromagnetic stud- ies aimed at investigating voluntary movement preparation and execution using the Bereitschaftspotential paradigm [9] are based on time domain analyses. However, evidence has accumulated that frequency domain analysis can provide important additional clues about how the human brain organizes a voluntary movement: after initial observations of Jasper and Penfield [8] and Chatrian et al. [1], Pfurtscheller and Aranibar [23] have shown that movement preparation is accompanied by a desynchronization and therefore amplitude depression of 'idling rhythms'. This event-related desynchronization (ERD, after Pfurtscheller and Aranibar [23]) lasts during voluntary movement prepa- ration and execution. ERD was shown to better differenti- ate between controls and Parkinsonians [3] and between * Corresponding author. Fax: +49 (731) 270-6619; E-mail: feige@ ukl.uni.freiburg.de younger and older subjects [4] than features of the Bere- itschaftspotential. Furthermore, it was shown by using intracortical recordings [33] that the idling rhythm genera- tors differ from the Bereitschaftspotential generators. ERD accompanying movement preparation and execution is usu- ally followed by an overshoot of spectral power above baseline values (event-related synchronization, ERS, after Pfurtscheller [21]). ERS is interpreted within the same deactivation framework to correspond to an extensive syn- chronization of the idling rhythm. While it is certainly true that high-amplitude oscillations visible in EEG and MEG correspond to cortical states of low differentiation and therefore of low information content (cf. [32,31]), it is possible that ERS could also reflect aspects of information processing rather than the mere inhibition of cortical areas. Previous studies often focused on spectral regions around discernible spectral power peaks. For example, in their recent study, Salmelin and Haft [26] have shown in spontaneous MEG signals that the two main frequency components (10 and 20 Hz) of the ~ rhythm, which is 0006-8993/96/$15.00 Copyright © 1996 Elsevier Science B.V. All rights reserved. PH S0006-8993(96)00648-8