Sleep Stage Classification Using Time-Frequency Spectra From Consecutive Multi-Time Points

Abstract

Sleep stage classification is an open challenge in the field of sleep research. Considering the relatively small size of datasets used by previous studies, in this paper we used the Sleep Heart Health Study dataset from the National Sleep Research Resource database. A long short-term memory (LSTM) network using a time-frequency spectra of several consecutive 30 s time points as an input was used to perform the sleep stage classification. Four classical convolutional neural networks (CNNs) using a time-frequency spectra of a single 30 s time point as an input were used for comparison. Results showed that, when considering the temporal information within the time-frequency spectrum of a single 30 s time point, the LSTM network had a better classification performance than the CNNs. Moreover, when additional temporal information was taken into consideration, the classification performance of the LSTM network gradually increased. It reached its peak when temporal information from three consecutive 30 s time points was considered, with a classification accuracy of 87.4% and a Cohen's Kappa coefficient of 0.8216. Compared with CNNs, our results indicate that for sleep stage classification, the temporal information within the data or the features extracted from the data should be considered. LSTM networks take this temporal information into account, and thus, may be more suitable for sleep stage classification.

Are the new mobile wireless EEG headsets reliable for the evaluation of musical pleasure?

“…We observed different results compared to our previous research that used the same data [ 35 ] and, despite a smaller sample, the only difference came from the unconventional referencing to P3 P4 for Epoc +. Considering those previous results, the increased significance with Epoc + in the present study does not mean that EGI is less performant, but only that the re-referencing largely shapes the signal.…”

“…These result are also concordant (for both devices) with physiological data from Rickard [ 53 ] and Salimpoor [ 54 ], suggesting that an increasing pleasure and the onset of chills produce an activation of the peripheral system with higher skin conductance responses/levels and higher heart rate levels. Our previous research did not demonstrate the relevance of approach withdrawal analysis with alpha asymmetry calculations with the EGI gold-standard setup [ 35 ]. Consequently, we chose to analyse only alpha and theta PSD measures.…”

“…The present results do not validate this hypothesis about theta oscillatory activity. Although the theta activity reflects the musical pleasure processing [ 35 , 42 , 45 – 47 ], this frequency band is also involved in other cognitive processes such as mental effort [ 55 , 56 ]. Consequently, whether theta is also involved in mental effort here should be investigated further.…”

“…In light of the results of our recent study on musical chills with EEG [ 35 ] and other research on musical pleasure [ 42 , 45 – 47 ], our main hypothesis was formulated for frontal activity. We expected that we would observe a gradual increase of theta power spectral density (PSD) values with the increase of emotional pleasure in the frontal/prefrontal areas with both devices.…”

Bliss in and Out of the Body: The (Extra)Corporeal Space Is Impervious to Social Pleasant Touch

“…Indeed, we traditionally conceive PpS as the space directly surrounding our body [ 36 , 37 ], which serves as a privileged interface between the body and the external world [ 36 ]. In accordance with this point of view, electro-physiological studies on monkey brain [ 38 ] and neuroscientific research in humans [ 39 , 40 ] showed that multisensory integration processes are gradually more powerful when they occur closer to the body. In particular, visuo-tactile interactions seem to be particularly sensitive to the spatial distance at which visual stimuli are positioned while the concurrent tactile input is delivered [ 39 , 41 , 42 ].…”