Ontogeny of delta activity during human sleep

Introduction. An easily accessible and inexpensive neurophysiological technique such as conventional electroencephalography may provide an accurate and generally applicable biomarker capable of differentiating dementia with Lewy bodies (DLB) from Alzheimer’s disease (AD) and Parkinson’s disease-associated dementia (PDD). Method. We carried out a retrospective visual analysis of resting-state electroencephalography (EEG) recording of 22 patients with a clinical diagnosis of 19 probable and 3 possible DLB, 22 patients with probable AD and 21 with PDD, matched for age, duration, and severity of cognitive impairment. Results. By using the grand total EEG scoring method, the total score and generalized rhythmic delta activity frontally predominant (GRDAfp) alone or, even better, coupled with a slowing of frequency of background activity (FBA) and its reduced reactivity differentiated DLB from AD at an individual level with an high accuracy similar to that obtained with quantitative EEG (qEEG). GRDAfp alone could also differentiate DLB from PDD with a similar level of diagnostic accuracy. AD differed from PDD only for a slowing of FBA. The duration and severity of cognitive impairment did not differ between DLB patients with and without GRDAfp, indicating that this abnormal EEG pattern should not be regarded as a disease progression marker. Conclusions. The findings of this investigation revalorize the role of conventional EEG in the diagnostic workup of degenerative dementias suggesting the potential inclusion of GRDAfp alone or better coupled with the slowing of FBA and its reduced reactivity, in the list of supportive diagnostic biomarkers of DLB.

Download Full-text

Capturing the Development of Internal Representations in a High-Performing Deep Network for Sleep Stage Classification

SN Computer Science ◽

10.1007/s42979-021-00697-3 ◽

2021 ◽

Vol 2 (4) ◽

Author(s):

Sarun Paisarnsrisomsuk ◽

Carolina Ruiz ◽

Sergio A. Alvarez

Keyword(s):

Input Data ◽

Single Channel ◽

Sleep Stage ◽

Quantitative Measure ◽

Classification Performance ◽

Sleep Stages ◽

Layer Depth ◽

Human Sleep ◽

Stage Classification ◽

Sleep Stage Classification

AbstractDeep neural networks can provide accurate automated classification of human sleep signals into sleep stages that enables more effective diagnosis and treatment of sleep disorders. We develop a deep convolutional neural network (CNN) that attains state-of-the-art sleep stage classification performance on input data consisting of human sleep EEG and EOG signals. Nested cross-validation is used for optimal model selection and reliable estimation of out-of-sample classification performance. The resulting network attains a classification accuracy of $$84.50 \pm 0.13\%$$ 84.50 ± 0.13 % ; its performance exceeds human expert inter-scorer agreement, even on single-channel EEG input data, therefore providing more objective and consistent labeling than human experts demonstrate as a group. We focus on analyzing the learned internal data representations of our network, with the aim of understanding the development of class differentiation ability across the layers of processing units, as a function of layer depth. We approach this problem visually, using t-Stochastic Neighbor Embedding (t-SNE), and propose a pooling variant of Centered Kernel Alignment (CKA) that provides an objective quantitative measure of the development of sleep stage specialization and differentiation with layer depth. The results reveal a monotonic progression of both of these sleep stage modeling abilities as layer depth increases.

Download Full-text