Decoding English Alphabet Letters Using EEG Phase Information

Abstract

Increasing evidence indicates that the phase pattern and power of the low frequency oscillations of brain electroencephalograms (EEG) contain significant information during the human cognition of sensory signals such as auditory and visual stimuli. Here, we investigate whether and how the letters of the alphabet can be directly decoded from EEG phase and power data. In addition, we investigate how different band oscillations contribute to the classification and determine the critical time periods. An English letter recognition task was assigned, and statistical analyses were conducted to decode the EEG signal corresponding to each letter visualized on a computer screen. We applied support vector machine (SVM) with gradient descent method to learn the potential features for classification. It was observed that the EEG phase signals have a higher decoding accuracy than the oscillation power information. Low-frequency theta and alpha oscillations have phase information with higher accuracy than do other bands. The decoding performance was best when the analysis period began from 180 to 380 ms after stimulus presentation, especially in the lateral occipital and posterior temporal scalp regions (PO7 and PO8). These results may provide a new approach for brain-computer interface techniques (BCI) and may deepen our understanding of EEG oscillations in cognition.

Converging intracortical signatures of two separated processing timescales in human early auditory cortex

“…In particular, phase yielded strongly significant (p<0.001) decoding results in 9/10 subjects for the hard discrimination, while power did so only in 4/10 subjects. Primacy of phase information over power information, especially for distinguishing between activity patterns corresponding to different sensory stimuli, is in accordance with several prior reports in auditory, visual and tactile modalities (Luo and Poeppel, 2007;Howard and Poeppel, 2010;Ng et al, 2013;Gross et al, 2013;Schyns et al, 2011;Ronconi et al, 2017;Wang et al, 2018;Baumgarten et al, 2015), in motor signals (Hammer et al, 2013), as well as in more complex paradigms involving working memory and decision making (Rizzuto et al, 2003;Lopour et al, 2013). In our dataset and for the binary decoding analyses we considered, the combination of power and phase did not convey additional information beyond the one conveyed by the most informative between power and phase ( Fig.…”

Temporal variabilities provide additional category-related information in object category decoding: a systematic comparison of informative EEG features

“…1-4 Hz). This finding was later investigated by another group using Hilbert transform, but found the information in the theta frequency band (4-8 Hz; Wang et al, 2018). Other studies found that signal power contained significant category-related information (Rupp et al, 2017;Majima et al, 2014;Miyakawa et al, 2018).…”

“…Later studies used Linear Discriminant Analysis (LDA) classifiers to discriminate up to four object categories utilizing the information content of those ERP components (Wang et al, 2012), which were later fused to improve previous decoding accuracies (Qin et al, 2016). However, these studies and others (Taghizadeh-Sarabi et al, 2015;Torabi et al, 2017;Wang et al, 2018) overlook the temporal dynamics of object encoding based on within-trial variations in the signals. To address this issue, researchers repeated the decoding procedure in short (e.g.…”

“…To address this issue, studies have investigated other features of brain activations such as phase (Behroozi et al, 2016;Watrous et al, 2015;Torabi et al, 2017;Wang et al, 2018), signal power across frequency bands (Rupp et al, 2017;Miyakawa et al, 2018;Behroozi et al, 2016), time-frequency features such as Wavelet coefficients Taghizadeh-Sarabi et al, 2015), inter-electrode correlation (Majima et al, 2014;Karimi-Rouzbahani et al, 2017a), non-linear statistical features (Joshi et al, 2018;Torabi et al, 2017;Stam, 2005). Behroozi et al, (2016) decoded object categories using signal phase patterns in the delta frequency band (i.e.…”

Decoding Imagined and Spoken Phrases From Non-invasive Neural (MEG) Signals

“…While a number of speech decoding studies have been conducted using EEG recently such as for classification of imagined syllables (D'Zmura et al, 2009;Brigham and Vijaya Kumar, 2010;Deng et al, 2010), isolated phonemes (Chi and John, 2011;Leuthardt et al, 2011;Zhao and Rudzicz, 2015;Yoshimura et al, 2016), alphabets (Wang et al, 2018), or words (Porbadnigk et al, 2009;Nguyen et al, 2017;Rezazadeh Sereshkeh et al, 2017), the decoding performances have been intermediate, e.g., 63.45% for a binary (yes/no) classification (Rezazadeh Sereshkeh et al, 2017) or 35.68% for five vowel classification (Cooney et al, 2019a). There are inherent disadvantages in using EEG that may have contributed to the difficulty in attaining high decoding performance.…”