Medically relevant criteria used in EEG compression for improved post-compression seizure detection

Abstract

Biomedical signals aid in the diagnosis of different disorders and abnormalities. When targeting lossy compression of such signals, the medically relevant information that lies within the data should maintain its accuracy and thus its reliability. In fact, signal models that are inspired by the biophysical properties of the signals at hand allow for a compression that preserves more naturally the clinically significant features of these signals. In this paper, we illustrate this through the example of EEG signals; more specifically, we analyze three specific lossy EEG compression schemes. These schemes are based on signal models that have different degrees of reliance on signal production and physiological characteristics of EEG. The resilience of these schemes is illustrated through the performance of seizure detection post compression.

Lossy compression techniques for EEG signals

“…Compression of EEG signals encourages the research and development of emerging medical applications such as telemedicine and AEEG system to enhance health care system and human life quality. The lossy compression features a promising technique for EEG signal based on its benefit in high CR especially in some special application in field of automatic detection of disease such as seizure and epilepsy [7,[35][36][37]. It was shown that SPIHT displays a clear advantage, achieving CR of 120 without impact on seizure detection [14].…”

“…It was proved that the seizure detection can achieve an accuracy up to 95% with PRD from 0 to 60% in the conscientious research for seizure detection by REACT while applying the reconstruction EEG signal after compressed by comparing the detected seizure with the annotation seizure signal in the Freiburg database [20]. Daou studied the effectiveness of three specific lossy compression algorithms (Dipole based, dictionary based and 2D SPIHT) on seizure detection by TP which is represented for the period of one minute or more of overlap occurs between a seizure section in the compressed file and a seizure section in the original file and claimed that a maximum value of TP of 100% was achievable [36]. The other works [7,14,27] also mentioned the possibility of applying lossy compressed for EEG signal, especially in seizure detection.…”