Driver drowsiness is associated with altered facial thermal patterns: Machine learning insights from a thermal imaging approach.

Driver drowsiness is a significant factor in road accidents. Thermal imaging has emerged as an effective tool for detecting drowsiness by enabling the analysis of facial thermal patterns. However, it is not clear which facial areas are most affected and correlate most strongly with drowsiness. This study examines the variations and importance of various facial areas and proposes an approach for detecting driver drowsiness. Twenty participants underwent tests in a driving simulator, and temperature changes in various facial regions were measured. The random forest method was employed to evaluate the importance of each facial region. The results revealed that temperature changes in the nasal area exhibited the highest value, while the eyes had the most correlated changes with drowsiness. Furthermore, drowsiness was classified with an accuracy of 88 % utilizing thermal variations in the facial region identified as the most important regions by the random forest feature importance model. These findings provide a comprehensive overview of facial thermal imaging for detecting driver drowsiness and introduce eye temperature as a novel and effective measure for investigating cognitive activities.

Beat-to-beat cardiac repolarization lability increases during hypoxemia and arousals in obstructive sleep apnea patients.

Obstructive sleep apnea (OSA) is associated with the progression of cardiovascular diseases, arrhythmias, and sudden cardiac death (SCD). However, the acute impacts of OSA and its consequences on heart function are not yet fully elucidated. We hypothesized that desaturation events acutely destabilize ventricular repolarization, and the presence of accompanying arousals magnifies this destabilization. Ventricular repolarization lability measures, comprising heart rate corrected QT (QTc), short-time-variability of QT (STVQT), and QT variability index (QTVI), were calculated before, during, and after 20,955 desaturations from lead II electrocardiography signals of 492 patients with suspected OSA (52% men). Variations in repolarization parameters were assessed during and after desaturations, both with and without accompanying arousals, and groupwise comparisons were performed based on desaturation duration and depth. Regression analyses were used to investigate the influence of confounding factors, comorbidities, and medications. The standard deviation (SD) of QT, mean QTc, SDQTc, and STVQT increased significantly (P < 0.01), whereas QTVI decreased (P < 0.01) during and after desaturations. The changes in SDQT, mean QTc, SDQTc, and QTVI were significantly amplified (P < 0.01) in the presence of accompanying arousals. Desaturation depth was an independent predictor of increased SDQTc (ß = 0.405, P < 0.01), STVQT (ß = 0.151, P < 0.01), and QTVI (ß = 0.009, P < 0.01) during desaturation. Desaturations cause acute changes in ventricular repolarization, with deeper desaturations and accompanying arousals independently contributing to increased ventricular repolarization lability. This may partially explain the increased risk of arrhythmias and SCD in patients with OSA, especially when the OSA phenotype includes high hypoxic load and fragmented sleep.NEW & NOTEWORTHY Nocturnal desaturations are associated with increased ventricular repolarization lability. Deeper desaturations with accompanying arousals increase the magnitude of alterations, independent of confounding factors, comorbidities, and medications. Changes associated with desaturations can partially explain the increased risk of arrhythmias and sudden cardiac death in patients with OSA, especially in patients with high hypoxic load and fragmented sleep. This highlights the importance of detailed electrocardiogram analytics for patients with OSA.

Inter-sleep stage variations in corrected QT interval differ between obstructive sleep apnea patients with and without stroke history.

Obstructive sleep apnea (OSA) is related to the progression of cardiovascular diseases (CVD); it is an independent risk factor for stroke and is also prevalent post-stroke. Furthermore, heart rate corrected QT (QTc) is an important predictor of the risk of arrhythmia and CVD. Thus, we aimed to investigate QTc interval variations in different sleep stages in OSA patients and whether nocturnal QTc intervals differ between OSA patients with and without stroke history. 18 OSA patients (apnea-hypopnea index (AHI)≥15) with previously diagnosed stroke and 18 OSA patients (AHI≥15) without stroke history were studied. Subjects underwent full polysomnography including an electrocardiogram measured by modified lead II configuration. RR, QT, and QTc intervals were calculated in all sleep stages. Regression analysis was utilized to investigate possible confounding effects of sleep stages and stroke history on QTc intervals. Compared to patients without previous stroke history, QTc intervals were significantly higher (ß = 34, p<0.01) in patients with stroke history independent of age, sex, body mass index, and OSA severity. N3 sleep (ß = 5.8, p<0.01) and REM sleep (ß = 2.8, p<0.01) increased QTc intervals in both patient groups. In addition, QTc intervals increased progressively (p<0.05) towards deeper sleep in both groups; however, the magnitude of changes compared to the wake stage was significantly higher (p<0.05) in patients with stroke history. The findings of this study indicate that especially in deeper sleep, OSA patients with a previous stroke have an elevated risk for QTc prolongation further increasing the risk for ventricular arrhythmogenicity and sudden cardiac death.

Multi-Level Classification of Driver Drowsiness by Simultaneous Analysis of ECG and Respiration Signals Using Deep Neural Networks.

The high number of fatal crashes caused by driver drowsiness highlights the need for developing reliable drowsiness detection methods. An ideal driver drowsiness detection system should estimate multiple levels of drowsiness accurately without intervening in the driving task. This paper proposes a multi-level drowsiness detection system by a deep neural network-based classification system using a combination of electrocardiogram and respiration signals. The proposed method is based on a combination of convolutional neural networks (CNNs) and long short-term memory (LSTM) networks for classifying drowsiness by concurrently using heart rate variability (HRV), power spectral density of HRV, and respiration rate signal as inputs. Two models, a CNN-based model and a hybrid CNN-LSTM-based model were used for multi-level classifications. The performance of the proposed method was evaluated on experimental data collected from 30 subjects in a simulated driving environment. The performance and the results of both models are presented and compared. The best performance for both three-level and five-level drowsiness classifications was achieved by the CNN-LSTM model. The results indicate that the three-level and five-level classifications of drowsiness can be achieved with 91 and 67% accuracy, respectively.