Corticomuscular and intermuscular coherence (CMC, IMC) reflect connectivity between neuronal activity in the motor cortex measured by electroencephalography (EEG) and muscular activity measured by electromyography (EMG), or between activity in different muscles, respectively. There is an ongoing debate on the appropriateness of EMG rectification prior to coherence estimation. This work examines the effects of EMG rectification in CMC and IMC estimation in 20 spinocerebellar ataxia type 2 (SCA2) patients, 16 prodromal SCA2 gene mutation carriers, and 26 healthy controls during a repetitive upper or lower limb motor task. Coherence estimations were performed using the non-rectified raw EMG signal vs. the rectified EMG signal. EMG rectification decreases the level of significance of lower beta-frequency band CMC and IMC values in SCA2 patients and prodromal SCA2 mutation carriers vs. healthy controls, and also results in overall lower coherence values. EMG rectification is detrimental for beta-frequency band CMC and IMC estimation. One likely reason for this effect is distortion of coherence estimation in high-frequency signals, where the level of amplitude cancelation is high.