Optimization of sigma amplitude threshold in sleep spindle detection

Abstract

Sleep spindles are transient EEG waveforms of non‐rapid eye movement sleep. There is considerable intersubject variability in spindle amplitudes. The problem in automatic spindle detection has been that, despite this fact, a fixed amplitude threshold has been used. Selection of the spindle detection threshold value is critical with respect to the sensitivity of spindle detection. In this study a method was developed to estimate the optimal recording‐specific threshold value for each all‐night recording without any visual scorings. The performance of the proposed method was validated using four test recordings each having a very different number of visually scored spindles. The optimal threshold values for the test recordings could be estimated well. The presented method seems very promising in providing information about sleep spindle amplitudes of individual all‐night recordings.

Sleep spindle detection through amplitude–frequency normal modelling

“…However, some comparisons were possible with works that published at least two statistic indicators common to the ones chosen in this paper. Huupponen et al (2000a) proposed an algorithm with a true positive rate ranging between 73.4% and 84.9% and a false positive rate ranging between 2.3% and 5.4%, hence with performance similar to our algorithm for the first study and slightly more sensitive and less selective for the second study. However, they evaluated their algorithm using the following criterion: "one second before or after a scored spindle or 1 s after a false-positive, no falsepositives were counted."…”

“…However, they evaluated their algorithm using the following criterion: "one second before or after a scored spindle or 1 s after a false-positive, no falsepositives were counted." If such a criterion was used to evaluate our algorithm, it would have shown a 4.3% false positive rate for the first study and 1.7% for the second one, hence tending to be lower than the one from Huupponen et al (2000a). The same group proposed another algorithm (Huupponen et al, 2007) with a true positive rate ranging between 51.2% and 86.5% and a false positive rate ranging between 26.4% and 46.0%.…”

“…Recent ones are based on methods that include fuzzy logic (Huupponen et al, 2000a(Huupponen et al, , 2003, neural network (Shimada et al, 2000;Huupponen et al, 2000b;Acir and Güzeliş , 2004;Ventouras et al, 2005;Güneş et al, 2011), bandpass filter (Clemens et al, 2005;Huupponen et al, 2007), fast time frequency transform (Knoblauch et al, 2003a(Knoblauch et al, , 2003b, Fourier transform (Huupponen et al, 2007;Duman et al, 2009), wavelet transform (Duman et al, 2009), Gabor transform (Schönwald et al, 2003) and matching pursuit (Durka et al, 2005;Schönwald et al, 2006;Ktonas et al, 2009).…”

“…Therefore, the 'classical' 12-14 Hz spindle definition is believed to be too narrow (Jankel and Niedermeyer, 1985). The difficulty in finding the optimum frequency bounds has produced a large number of proposed values, among them: 11.5-15 Hz (Fish et al, 1988), 11.5-16 Hz (Zeitlhofer et al, 1997), 11-15 Hz (Ktonas et al, 2009), 11-16 Hz (Clemens et al, 2005, 10.5-16 Hz (Ventouras et al, 2005;Huupponen et al, 2007), and 10-16 Hz (Zygierewicz et al, 1999;Huupponen et al, 2000a;Estévez et al, 2002). Beside the often cited 12-14 Hz frequency range proposed by the National Institute of Neurological Diseases and Blindness of the U.S. Department of Health, Education, and Welfare (Rechtschaffen and Kales, 1968), various organizations have suggested other values to score spindles: 11-16 Hz by the American Academy of Sleep Medicine (Iber et al, 2007), 11-15 Hz by the International Federation of Clinical Neurophysiology (Noachtar et al, 1999), and 12-16 Hz by the Japanese Society of Sleep Research (Hori et al, 2001).…”

“…Indeed, there is a large variability in both amplitude (Huupponen et al, 2000a;Clemens et al, 2005) and frequency (Zeitlhofer et al, 1997) across subjects and in both amplitude and frequency across derivations (Zeitlhofer et al, 1997). Therefore, it seems difficult to design a spindle detection algorithm without adapting to this variability.…”