Background: Delirium in hospitalized patients is a syndrome of acute brain dysfunction. Diagnostic (International Classification of Diseases [ICD]) codes are often used in studies using electronic health records (EHRs), but they are inaccurate. Objective: We sought to develop a more accurate method using natural language processing (NLP) to detect delirium episodes on the basis of unstructured clinical notes. Methods: We collected 1.5 million notes from >10,000 patients from among 9 hospitals. Seven experts iteratively labeled 200,471 sentences. Using these, we trained three NLP classifiers: Support Vector Machine, Recurrent Neural Networks, and Transformer. Testing was performed using an external data set. We also evaluated associations with delirium billing (ICD) codes, medications, orders for restraints and sitters, direct assessments (Confusion Assessment Method [CAM] scores), and in-hospital mortality. F1 scores, confusion matrices, and areas under the receiver operating characteristic curve (AUCs) were used to compare NLP models. We used the φ coefficient to measure associations with other delirium indicators. Results: The transformer NLP performed best on the following parameters: micro F1=0.978, macro F1=0.918, positive AUC=0.984, and negative AUC=0.992. NLP detections exhibited higher correlations (φ) than ICD codes with deliriogenic medications (0.194 vs 0.073 for ICD codes), restraints and sitter orders (0.358 vs 0.177), mortality (0.216 vs 0.000), and CAM scores (0.256 vs –0.028). Conclusions: Clinical notes are an attractive alternative to ICD codes for EHR delirium studies but require automated methods. Our NLP model detects delirium with high accuracy, similar to manual chart review. Our NLP approach can provide more accurate determination of delirium for large-scale EHR-based studies regarding delirium, quality improvement, and clinical trails.

Alterations in sleep spindles have been linked to cognitive impairment. This finding has contributed to a growing interest in identifying sleep-based biomarkers of cognition and neurodegeneration, including sleep spindles. However, flexibility surrounding spindle definitions and algorithm parameter settings present a methodological challenge. The aim of this study was to characterize how spindle detection parameter settings influence the association between spindle features and cognition and to identify parameters with the strongest association with cognition. Adult patients (n = 167, 49 ± 18 years) completed the NIH Toolbox Cognition Battery after undergoing overnight diagnostic polysomnography recordings for suspected sleep disorders. We explored 1000 combinations across seven parameters in Luna, an open-source spindle detector, and used four features of detected spindles (amplitude, density, duration, and peak frequency) to fit linear multiple regression models to predict cognitive scores. Spindle features (amplitude, density, duration, and mean frequency) were associated with the ability to predict raw fluid cognition scores (r = 0.503) and age-adjusted fluid cognition scores (r = 0.315) with the best spindle parameters. Fast spindle features generally showed better performance relative to slow spindle features. Spindle features weakly predicted total cognition and poorly predicted crystallized cognition regardless of parameter settings. Our exploration of spindle detection parameters identified optimal parameters for studies of fluid cognition and revealed the role of parameter interactions for both slow and fast spindles. Our findings support sleep spindles as a sleep-based biomarker of fluid cognition.

Backgrounds: Sleep disturbances are prevalent among elderly individuals. While polysomnography (PSG) serves as the gold standard for sleep monitoring, its extensive setup and data analysis procedures impose significant costs and time constraints, thereby restricting the long-term application within the general public. Our laboratory introduced an innovative biomarker, utilizing artificial intelligence algorithms applied to PSG data to estimate brain age (BA), a metric validated in cohorts with cognitive impairments. Nevertheless, the potential of exercise, which has been a recognized means of enhancing sleep quality in middle-aged and older adults to reduce BA, remains undetermined. Methods: We conducted an exploratory study to evaluate whether 12 weeks of moderate-intensity exercise can improve cognitive function, sleep quality, and the brain age index (BAI), a biomarker computed from overnight sleep electroencephalogram (EEG), in physically inactive middle-aged and older adults. Home wearable devices were used to monitor heart rate and overnight sleep EEG over this period. The NIH Toolbox Cognition Battery, in-lab overnight polysomnography, cardiopulmonary exercise testing, and a multiplex cytokines assay were employed to compare pre- and post-exercise brain health, exercise capacity, and plasma proteins. Results: In total, 26 participants completed the initial assessment and exercise program, and 24 completed all procedures. Data are presented as mean [lower 95% CI of mean, upper 95% CI of mean]. Participants significantly increased maximal oxygen consumption (Pre: 21.11 [18.98, 23.23], Post 22.39 [20.09, 24.68], mL/kg/min; effect size: −0.33) and decreased resting heart rate (Pre: 66.66 [63.62, 67.38], Post: 65.13 [64.25, 66.93], bpm; effect size: −0.02) and sleeping heart rate (Pre: 64.55 [61.87, 667.23], Post: 62.93 [60.78, 65.09], bpm; effect size: −0.15). Total cognitive performance (Pre: 111.1 [107.6, 114.6], Post: 115.2 [111.9, 118.5]; effect size: 0.49) was significantly improved. No significant differences were seen in BAI or measures of sleep macro- and micro-architecture. Plasma IL-4 (Pre: 0.24 [0.18, 0.3], Post: 0.33 [0.24, 0.42], pg/mL; effect size: 0.49) was elevated, while IL-8 (Pre: 5.5 [4.45, 6.55], Post: 4.3 [3.66, 5], pg/mL; effect size: −0.57) was reduced. Conclusions: Cognitive function was improved by a 12-week moderate-intensity exercise program in physically inactive middle-aged and older adults, as were aerobic fitness (VO₂max) and plasma cytokine profiles. However, we found no measurable effects on sleep architecture or BAI. It remains to be seen whether a study with a larger sample size and more intensive or more prolonged exercise exposure can demonstrate a beneficial effect on sleep quality and brain age.

Sleep underpins cognition, disease prevention, and overall brain health, yet objective, integrative biomarkers of brain health remain lacking. We hypothesized that overnight sleep electroencephalography (EEG) could provide a substrate for such a biomarker. We asked whether a newly developed, end-to-end, data-driven deep learning framework for sleep EEG can learn a latent representation of brain health and distill it into a single score relevant to cognition, disease status, and mortality. We analyzed 36,000 polysomnography recordings from 27,000 subjects from six cohorts. EEG data were represented as one-dimensional time series or a two-dimensional time-frequency spectrogram. A multitask deep neural network, trained end-to-end without expert-defined features, learned a 1024-dimensional brain health latent space and jointly predicted cognitive performance, disease status, and sleep metrics. The latent representation was additionally distilled into a single brain health score. We compared performance with demographic baselines, conventional EEG metrics (e.g., rapid eye movement fraction, spindle density), and classic multivariate machine learning approaches. The deep learning–derived brain health scores consistently surpassed demographic and expert-defined EEG feature models. For cognitive outcomes, correlations (r) rose from small (demographic-only) to moderate (up to r=0.40), while disease classification areas under the receiver operator curve improved from 0.50–0.55 at baseline to 0.65–0.75. In age-adjusted Cox models, a one-standard-deviation increase in the brain health score was associated with a 31%–35% reduced risk of mortality (hazard ratio 0.65 to 0.69; P<0.0001), topping conventional EEG metrics. Gains over classic machine learning, plus latent space visualization, indicated that both established physiological markers and novel EEG features drove enhanced performance. A multitask, end-to-end deep learning approach generated an interpretable, sleep-derived brain health biomarker. By modeling cognition, disease, and mortality, this framework provides a robust index of brain health and may be extended to additional modalities, further enhancing its clinical utility. (Funded by the National Institutes of Health and others.)

Background and Purpose: Reports have suggested that severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) causes neurologic manifestations including encephalopathy and seizures. However, there has been relatively limited electrophysiology data to contextualize these specific concerns and to understand their associated clinical factors. Our objective was to identify EEG abnormalities present in patients with SARS-CoV-2, and to determine whether they reflect new or preexisting brain pathology. Methods: We studied a consecutive series of hospitalized patients with SARS-CoV-2 who received an EEG, obtained using tailored safety protocols. Data from EEG reports and clinical records were analyzed to identify EEG abnormalities and possible clinical associations, including neurologic symptoms, new or preexisting brain pathology, and sedation practices. Results: We identified 37 patients with SARS-CoV-2 who underwent EEG, of whom 14 had epileptiform findings (38%). Patients with epileptiform findings were more likely to have preexisting brain pathology (6/14, 43%) than patients without epileptiform findings (2/23, 9%; p = 0.042). There were no clear differences in rates of acute brain pathology. One case of nonconvulsive status epilepticus was captured, but was not clearly a direct consequence of SARS-CoV-2. Abnormalities of background rhythms were common, as may be seen in systemic illness, and in part associated with recent sedation (p = 0.022). Conclusions: Epileptiform abnormalities were common in patients with SARS-CoV-2 referred for EEG, but particularly in the context of preexisting brain pathology and sedation. These findings suggest that neurologic manifestations during SARS-CoV-2 infection may not solely relate to the infection itself, but rather may also reflect patients’ broader, preexisting neurologic vulnerabilities.

Objectives: We aimed to quantify the effects of advanced meditation on brain electrical activity during sleep. This investigation addresses the need for objective neurophysiological measures of meditation’s potential impact on brain aging and health. Method: This study was a single-site, prospective cohort study (conducted August 25, 2021, through September 26, 2021) of meditators attending the “Samyama Sadhana” retreat (September 1–5, 2021). Two healthy comparison groups and four comparison groups with varying degrees of age-related brain pathology are included. Using overnight electroencephalography, physiological measures of brain age were derived and subtracted from chronological age, measuring the deviation of apparent brain age from chronological age. Results: Thirty-four participants completed the study (average age = 38 years; 36% female). Estimated brain age index after adjustment by matching: meditators (n = 34), − 5.9 years (SE = 0.94 years, t-test p < 0.001); Dreem healthy controls (n = 1077), − 0.24 (0.61, p < 0.001); Massachusetts General Hospital (MGH) healthy controls (n = 112), 0.55 (0.92, p < 0.05); MGH “no dementia” (n = 7618), 2.4 (0.094, reference cohort for t-test); MGH “symptomatic” (n = 697), 2.0 (0.33, p > 0.05); MGH “mild cognitive impairment (MCI)”(n = 205), 8.8 (2.8, p < 0.05); and MGH “dementia” (n = 153), 10.5 (2.8, p < 0.01). Conclusions: Long-term meditators exhibit lower brain age relative to matched control groups. This study suggests that advanced meditation enhances brain health. Preregistration: This study was not preregistered.