The cerebral oscillatory network of voluntary tremor

Abstract

It has recently been shown that resting tremor in Parkinson's disease is associated with oscillatory neural coupling in an extensive cerebral network comprising a cerebello-diencephalic-cortical loop and cortical motor, somatosensory and posterior parietal areas contralateral to the tremor hand. The aim of the present study was to investigate whether this oscillatory brain network exclusively reflects a pathophysiological state in parkinsonian resting tremor or whether it constitutes a fundamental feature of physiological motor control. We investigated cerebro-muscular and cerebro-cerebral coupling in 11 healthy subjects imitating typical antagonistic parkinsonian tremor. We recorded brain activity with a 122-channel whole-head neuromagnetometer and surface EMGs of the forearm extensor. Analysis of cerebro-muscular and cerebro-cerebral coherence revealed oscillatory coupling in the same brain structures that comprise the oscillatory network of parkinsonian resting tremor. Interestingly, similar to parkinsonian resting tremor, cerebro-cerebral coherences often showed a significant peak at twice the simulated tremor frequency. The most striking differences between parkinsonian patients, as investigated in a previous study and healthy subjects imitating the antagonistic resting tremor were a reduction of the coupling between primary sensorimotor cortex and a diencephalic structure--most likely the thalamus--and an enhancement of the coupling between premotor and primary sensorimotor cortex. Our results indicate that the coupling of oscillatory activity within a cerebello-diencephalic-cortical loop constitutes a basic feature of physiological motor control. Thus, our data are consistent with the hypothesis that parkinsonian resting tremor involves oscillatory cerebro-cerebral coupling in a physiologically pre-existing network.

Figure 1. Analysis of cerebro-muscular coherence in a representative subject imitating parkinsonian resting tremor with the dominant right hand

A, left, traces of surface EMG activity of right extensor digitorum (EDC) (upper trace) and of right flexor digitorum longus (FDL) muscle (lower trace). The EMG was highpass filtered at 60 Hz and rectified. Regular EMG bursts occur at the frequency of the imitated tremor of 4 Hz. Right, power spectral activity of the EDC. We found two main peaks: at tremor frequency at about 4 Hz and a smaller peak at double tremor frequency. B, coherence between the EMG of the imitating hand and the 122 MEG sensors as viewed from above. Significant coherence was observed via the sensors covering the sensorimotor cortex contralateral to the imitating hand. The inset shows the sensor with the highest coherence. Coherence between EMG and MEG was observed at tremor frequency and was even stronger at double tremor frequency. C, left, localization of cerebro-muscular coherence at tremor frequency to right EDC as revealed with DICS in the individual MRI scan. In all subjects the source with the strongest coherence to the extensor digitorum muscle was localized in the sensorimotor hand area. Right, coherence between EMG and the S1/M1 source at the tremor and at double tremor frequency. Coherence was strongest at double tremor frequency. The dashed line indicates the 95% confidence level of coherence.

Figure 2. Localization of cerebral sources and coherence spectra to right EDC muscle in another representative subject imitating the PD resting tremor with the right hand

Note, that only the S1/M1 source was localized with respect to the EDC. All other sources were detected with reference to the S1/M1 source. Coherent activity was found in the contralateral primary sensorimotor cortex (A), premotor cortex (B), ipsilateral cerebellum (C), diencephalon (D), secondary somatosensory cortex (E) and posterior parietal cortex (F). All sources were significantly coherent to the right EDC with the exception of the thalamic source where, in this subject, discernible spectral coherence peaks at tremor and twice the tremor frequency failed to reach significance. Note, that strength of source localization is colour coded. Red represents stronger coherent activation whereas blue indicates less coherence. Dotted lines indicate the 95% confidence level of coherence. Note, that significant peaks were found at tremor frequency and, in part, even stronger at double tremor frequency (A).

Figure 3. Cerebro-cerebral coherence between different brain areas in one representative subject

Dotted lines indicate the 95% confidence level of coherence. Note, that coupling does not always occur exactly at tremor frequency or at double tremor frequency. Frequency shifts are not characteristic for coupling between specific areas. They might be due to the calculation resolution of 0.98 Hz, which might lead to shifts up to ± 2 Hz.