Improving the detection of auditory steady-state responses near 80 Hz using multiple magnitude-squared coherence and multichannel electroencephalogram

Methodological considerations when measuring and analyzing auditory steady-state responses with multi-channel EEG

“…Both the multi-channel Hotelling's T 2 (e.g. Mijares et al, 2013 ; Vanheusden et al, 2019 ) and MMSC ( da Silva Eloi et al, 2018 ; Felix et al, 2018 ) methods have been used in recent multi-channel EEG studies of ASSR, and both aim to provide an overall estimate of ASSR strength with phase differences between channels corrected by the estimated covariance matrix.…”

“…This assumption may not always hold, as the scalp ASSR likely reflects a mixture of multiple neural generators with different phases ( Bidelman, 2015 , 2018 ; Chandrasekaran and Kraus, 2010 ; Coffey et al, 2019 ). To accommodate such phase inconsistencies, more sophisticated measures have been developed, such as the multiple magnitude-squared coherence (MMSC) ( da Silva Eloi et al, 2018 ), complex principal component analysis (cPCA) ( Bharadwaj and Shinn-Cunningham, 2014 ; Lu et al, 2020 ), and the Hotelling's T 2 ( Picton et al, 1987 ). Finally, an alternative approach to accommodate the possibility of unequal phases is to independently compute ASSR strength within each single channel and to simply summarize the output from all electrodes by averaging or computing other statistics such as root-mean-square (RMS) value.…”

Suppressing the Spikes in Electroencephalogram via an Iterative Joint Singular Spectrum Analysis and Low-Rank Decomposition Approach

“…subject to Y = S + E. (6b) Figure 1 shows the flowchart of the proposed method. Since most of the information of the electroencephalogram is found in the δ wave band (0-4 Hz), the θ wave band (4-8 Hz), the α wave band (8)(9)(10)(11)(12), the β wave band and the γ wave band (30-44 Hz), these frequency bands are the most important frequency bands for performing the electroencephalograph analysis [1][2][3]. To suppress the noise outside these signal bands, a simple lowpass filtering is applied to the electroencephalogram.…”

“…Electroencephalogram is a brain signal that reflects the activities of the human body such as the splendid, the sleeping quality, the emotion level and the epileptic disorder. By studying the electroencephalogram, many related diseases can be diagnosed effectively [1][2][3]. However, the quality of the acquired signal is usually very poor-this may be due to the sensor limitations, the background environmental noise and the human factors.…”