Cortical correlates of the basic and first harmonic frequency of Parkinsonian tremor

Abstract

Objective

It has been hypothesized that the basic and first harmonic frequency of Parkinsonian tremor are somewhat independent oscillations the biological basis of which remains unclear.

Methods

We recorded 64-channel EEG in parallel with EMG of the forearm muscles most affected by rest tremor in 21 PD patients. EMG power spectrum, corticomuscular coherence spectra and EEG power spectra for each EEG electrode were calculated. The dynamics of the coherence and relative EMG and EEG power at the basic (tremor) frequency were calculated by a sliding, overlapping window analysis. Corticomuscular delays and direction of interaction were analysed by the maximizing coherence method for narrow band signals.

Results

The contralateral EEG electrodes with maximal coherence were different for the basic and first harmonic frequency. The dynamical coherence curves showed non-parallel time courses for the two frequencies. The mean EEG-EMG and EMG-EEG delays were all around 15–20 ms but significantly longer for the first harmonic than for the basic frequency.

Conclusions

Our data indicate different cortical representations and corticomuscular interaction of the basic and first harmonic frequencies of Parkinsonian tremor.

Significance

Separate central generators seem to contribute to the tremor via different pathways. Further studies on this complex tremor network are warranted.

Introduction

Classical Parkinsonian resting tremor is very regular with a typical frequency between 3 and 6 Hz. However, in the power spectrum of the underlying rhythmic muscle activity we typically find a peak not only at the tremor frequency but also at double this frequency (Spieker et al., 1995, Deuschl et al., 1996, Milanov, 2000). This higher frequency peak is termed first higher harmonic, as it is a well known physical phenomenon that rhythmic processes with nonlinear or asymmetric wave forms produce peaks not only at their actual frequency (basic frequency) but also harmonic peaks at integer multiples of the basic frequency (Deuschl et al., 2000). In such cases the higher harmonic peaks can be considered to reflect a part of the same rhythmic process at the basic frequency. As wave form analyses have shown asymmetries in Parkinsonian tremor (Deuschl et al., 1995) its first higher harmonic peak has usually been interpreted in line with this physical rule. But this interpretation has been questioned recently. On the basis of MEG–EMG analyses it has been postulated that the first higher harmonic is the main corticospinal drive contributing to the peripheral tremor (Timmermann et al., 2003). A recent analysis in the time domain has lent further support to the view that the first harmonic peak may reflect an independent phenomenon (Sapir et al., 2003) rather than a mere effect of wave form asymmetry. However, the biological basis of such independent oscillations in Parkinsonian tremor remains obscure. One possible explanation would be separate central (cortical) generators as has been alluded to by (Volkmann et al., 1996). If this assumption was correct one would expect differing temporal and spatial patterns of the cortical correlates of the basic and higher harmonic frequencies in Parkinsonian tremor. We therefore examined the corticomuscular (EEG-EMG) coherence in such patients specifically looking for differences in the distribution of the coherences on the scalp and in their dynamics over time.