A Deep Learning Driven Simulation Analysis of the Emotional Profiles of Depression Based on Facial Expression Dynamics.

Objective: : Diagnosis and assessment of depression rely on scoring systems based on questionnaires, either self-reported by patients or administered by clinicians, and observation of patient facial expressions during the interviews plays a crucial role in making impressions in clinical settings. Deep learning driven approaches can assist clinicians in the course of diagnosis of depression by recognizing subtle facial expressions and emotions in depression patients. Methods: : Seventeen simulated patients who acted as depressed patients participated in this study. A trained psychiatrist structurally interviewed each participant with moderate depression in accordance with a prepared scenario and without depressive features. Interviews were video-recorded, and a facial emotion recognition algorithm was used to classify emotions of each frame. Results: : Among seven emotions (anger, disgust, fear, happiness, neutral, sadness, and surprise), sadness was expressed in a higher proportion on average in the depression-simulated group compared to the normal group. Neutral and fear were expressed in higher proportions on average in the normal group compared to the normal group. The overall distribution of emotions between the two groups was significantly different (p ＜ 0.001). Variance in emotion was significantly less in the depression-simulated group (p ＜ 0.05). Conclusion: : This study suggests a novel and practical approach to understand the emotional expression of depression patients based on deep learning techniques. Further research would allow us to obtain more perspectives on the emotional profiles of clinical patients, potentially providing helpful insights in making diagnosis of depression patients.

Prediction of Mortality in Surgical Intensive Care Unit Patients Using Machine Learning Algorithms.

Objective: Predicting prognosis of in-hospital patients is critical. However, it is challenging to accurately predict the life and death of certain patients at certain period. To determine whether machine learning algorithms could predict in-hospital death of critically ill patients with considerable accuracy and identify factors contributing to the prediction power. Materials and Methods: Using medical data of 1,384 patients admitted to the Surgical Intensive Care Unit (SICU) of our institution, we investigated whether machine learning algorithms could predict in-hospital death using demographic, laboratory, and other disease-related variables, and compared predictions using three different algorithmic methods. The outcome measurement was the incidence of unexpected postoperative mortality which was defined as mortality without pre-existing not-for-resuscitation order that occurred within 30 days of the surgery or within the same hospital stay as the surgery. Results: Machine learning algorithms trained with 43 variables successfully classified dead and live patients with very high accuracy. Most notably, the decision tree showed the higher classification results (Area Under the Receiver Operating Curve, AUC = 0.96) than the neural network classifier (AUC = 0.80). Further analysis provided the insight that serum albumin concentration, total prenatal nutritional intake, and peak dose of dopamine drug played an important role in predicting the mortality of SICU patients. Conclusion: Our results suggest that machine learning algorithms, especially the decision tree method, can provide information on structured and explainable decision flow and accurately predict hospital mortality in SICU hospitalized patients.

Severe COVID-19 Illness: Risk Factors and Its Burden on Critical Care Resources.

In South Korea, the first confirmed case of coronavirus 2019 (COVID-19) was detected on January 20, 2020. After a month, the number of confirmed cases surged, as community transmission occurred. The local hospitals experienced severe shortages in medical resources such as mechanical ventilators and extracorporeal membrane oxygenation (ECMO) equipment. With the medical claims data of 7,590 COVID-19 confirmed patients, this study examined how the demand for major medical resources and medications changed during the outbreak and subsequent stabilization period of COVID-19 in South Korea. We also aimed to investigate how the underlying diseases and demographic factors affect disease severity. Our findings revealed that the risk of being treated with a mechanical ventilator or ECMO (critical condition) was almost twice as high in men, and a previous history of hypertension, diabetes, and psychiatric diseases increased the risk for progressing to critical condition [Odds Ratio (95% CI), 1.60 (1.14-2.24); 1.55 (1.55-2.06); 1.73 (1.25-2.39), respectively]. Although chronic pulmonary disease did not significantly increase the risk for severity of the illness, patients with a Charlson comorbidity index score of ≥5 and those treated in an outbreak area had an increased risk of developing a critical condition [3.82 (3.82-8.15); 1.59 (1.20-2.09), respectively]. Our results may help clinicians predict the demand for medical resources during the spread of COVID-19 infection and identify patients who are likely to develop severe disease.

Causally linking neural dominance to perceptual dominance in a multisensory conflict.

When different senses are in conflict, one sense may dominate the perception of other sense, but it is not known whether the sensory cortex associated with the dominant modality exerts directional influence, at the functional brain level, over the sensory cortex associated with the dominated modality; in short, the link between sensory dominance and neuronal dominance is not established. In a task involving audio-visual conflict, using magnetoencephalography recordings in humans, we first demonstrated that the neuronal dominance - auditory cortex functionally influencing visual cortex - was associated with the sensory dominance - sound qualitatively altering visual perception. Further, we found that prestimulus auditory-to-visual connectivity could predict the perceptual outcome on a trial-by-trial basis. Subsequently, we performed an effective connectivity-guided neurofeedback electroencephalography experiment and showed that participants who were briefly trained to increase the neuronal dominance from auditory to visual cortex showed higher sensory, that is auditory, dominance during the conflict task immediately after the training. These results shed new light into the interactive neuronal nature of multisensory integration and open up exciting opportunities by enhancing or suppressing targeted mental functions subserved by effective connectivity.

Association Between Macronutrients Intake and Depression in the United States and South Korea.

Although the risk for depression appears to be related to daily dietary habits, how the proportion of major macronutrients affects the occurrence of depression remains largely unknown. This study aims to estimate the association between macronutrients (i.e., carbohydrate, protein, fat) and depression through national survey datasets from the United States and South Korea. Association between the prevalence of depression and each macronutrient was measured from 60,935 participants from the National Health and Nutrition Examination Survey (NHANES) and 15,700 participants from the South Korea NHANES (K-NHANES) databases. When the proportion of calories intake by protein increased by 10%, the prevalence of depression was significantly reduced both in the United States [Odds Ratio, OR (95% CI), 0.621 (0.530-0.728)] and South Korea [0.703 (0.397-0.994)]. An association between carbohydrate intake and the prevalence of depression was seen in the United States [1.194 (1.116-1.277)], but not in South Korea. Fat intake was not significantly associated with depression in either country. Subsequent analysis showed that the low protein intake groups had significantly higher risk for depression than the normal protein intake groups in both the United States [1.648 (1.179-2.304)] and South Korea [3.169 (1.598-6.286)]. In the daily diet of macronutrients, the proportion of protein intake is significantly associated with the prevalence of depression. These associations were more prominent in adults with insufficient protein intake, and the pattern of association between macronutrients and depression in Asian American and South Korean populations were similar. Our findings suggest that the proportion of macronutrients intake in everyday life may be related to the occurrence of depression.

Identifying Schizophrenia Using Structural MRI With a Deep Learning Algorithm.

OBJECTIVE: Although distinctive structural abnormalities occur in patients with schizophrenia, detecting schizophrenia with magnetic resonance imaging (MRI) remains challenging. This study aimed to detect schizophrenia in structural MRI data sets using a trained deep learning algorithm. METHOD: Five public MRI data sets (BrainGluSchi, COBRE, MCICShare, NMorphCH, and NUSDAST) from schizophrenia patients and normal subjects, for a total of 873 structural MRI data sets, were used to train a deep convolutional neural network. RESULTS: The deep learning algorithm trained with structural MR images detected schizophrenia in randomly selected images with reliable performance (area under the receiver operating characteristic curve [AUC] of 0.96). The algorithm could also identify MR images from schizophrenia patients in a previously unencountered data set with an AUC of 0.71 to 0.90. The deep learning algorithm's classification performance degraded to an AUC of 0.71 when a new data set with younger patients and a shorter duration of illness than the training data sets was presented. The brain region contributing the most to the performance of the algorithm was the right temporal area, followed by the right parietal area. Semitrained clinical specialists hardly discriminated schizophrenia patients from healthy controls (AUC: 0.61) in the set of 100 randomly selected brain images. CONCLUSIONS: The deep learning algorithm showed good performance in detecting schizophrenia and identified relevant structural features from structural brain MRI data; it had an acceptable classification performance in a separate group of patients at an earlier stage of the disease. Deep learning can be used to delineate the structural characteristics of schizophrenia and to provide supplementary diagnostic information in clinical settings.

Neural Correlates of Public Apology Effectiveness.

Apologizing is an effective interpersonal conflict resolution strategy, but whether, and if so how, organizations should issue public apologies after crises remains less clear. To assuage the fear of possible crisis reoccurrence, public apologies may be effective when they provide a comprehensive account of what happened and clarify actions taken by the company to address the problems. If this is so, public apologies may be most effective when the crisis source resides within the organization itself, suggesting that the company has control over it. In the current study, we first tested this hypothesis by presenting participants with multiple crisis scenarios (e.g., ignition failures in a new car model) followed by one of two written apologies: one stating that the crisis source was internal to and controllable by the organization, and the other external and uncontrollable. The internal-controllable (IC) public apology proved most effective. We then examined the neural basis of this public apology assessment and found that the frontal polar cortex appears to mediate the assessment of organizational control, and the angular gyrus uses the information for the apology assessment. Examination of complex social interactions, such as the public's reaction to corporate crises, helps to elucidate high-level brain function.

Identifying depression in the National Health and Nutrition Examination Survey data using a deep learning algorithm.

BACKGROUND: As depression is the leading cause of disability worldwide, large-scale surveys have been conducted to establish the occurrence and risk factors of depression. However, accurately estimating epidemiological factors leading up to depression has remained challenging. Deep-learning algorithms can be applied to assess the factors leading up to prevalence and clinical manifestations of depression. METHODS: Customized deep-neural-network and machine-learning classifiers were assessed using survey data from 19,725 participants from the NHANES database (from 1999 through 2014) and 4949 from the South Korea NHANES (K-NHANES) database in 2014. RESULTS: A deep-learning algorithm showed area under the receiver operating characteristic curve (AUCs) of 0.91 and 0.89 for detecting depression in NHANES and K-NHANES, respectively. The deep-learning algorithm trained with serial datasets (NHANES, from 1999 to 2012), predicted the prevalence of depression in the following two years of data (NHANES, 2013 and 2014) with an AUC of 0.92. Machine learning classifiers trained with NHANES could further predict depression in K-NHANES. There, logistic regression had the highest performance (AUC, 0.77) followed by deep learning algorithm (AUC, 0.74). CONCLUSIONS: Deep neural-networks managed to identify depression well from other health and demographic factors in both the NHANES and K-NHANES datasets. The deep-learning algorithm was also able to predict depression relatively well on new data set-cross temporally and cross nationally. Further research can delineate the clinical implications of machine learning and deep learning in detecting disease prevalence and progress as well as other risk factors for depression and other mental illnesses.

Classification of Suicide Attempts through a Machine Learning Algorithm Based on Multiple Systemic Psychiatric Scales.

Classification and prediction of suicide attempts in high-risk groups is important for preventing suicide. The purpose of this study was to investigate whether the information from multiple clinical scales has classification power for identifying actual suicide attempts. Patients with depression and anxiety disorders (N = 573) were included, and each participant completed 31 self-report psychiatric scales and questionnaires about their history of suicide attempts. We then trained an artificial neural network classifier with 41 variables (31 psychiatric scales and 10 sociodemographic elements) and ranked the contribution of each variable for the classification of suicide attempts. To evaluate the clinical applicability of our model, we measured classification performance with top-ranked predictors. Our model had an overall accuracy of 93.7% in 1-month, 90.8% in 1-year, and 87.4% in lifetime suicide attempts detection. The area under the receiver operating characteristic curve (AUROC) was the highest for 1-month suicide attempts detection (0.93), followed by lifetime (0.89), and 1-year detection (0.87). Among all variables, the Emotion Regulation Questionnaire had the highest contribution, and the positive and negative characteristics of the scales similarly contributed to classification performance. Performance on suicide attempts classification was largely maintained when we only used the top five ranked variables for training (AUROC; 1-month, 0.75, 1-year, 0.85, lifetime suicide attempts detection, 0.87). Our findings indicate that information from self-report clinical scales can be useful for the classification of suicide attempts. Based on the reliable performance of the top five predictors alone, this machine learning approach could help clinicians identify high-risk patients in clinical settings.

Transcranial Alternating Current Stimulation (tACS) Mechanisms and Protocols.

Perception, cognition and consciousness can be modulated as a function of oscillating neural activity, while ongoing neuronal dynamics are influenced by synaptic activity and membrane potential. Consequently, transcranial alternating current stimulation (tACS) may be used for neurological intervention. The advantageous features of tACS include the biphasic and sinusoidal tACS currents, the ability to entrain large neuronal populations, and subtle control over somatic effects. Through neuromodulation of phasic, neural activity, tACS is a powerful tool to investigate the neural correlates of cognition. The rapid development in this area requires clarity about best practices. Here we briefly introduce tACS and review the most compelling findings in the literature to provide a starting point for using tACS. We suggest that tACS protocols be based on functional brain mechanisms and appropriate control experiments, including active sham and condition blinding.

Changes in cerebral glucose metabolism after 3 weeks of noninvasive electrical stimulation of mild cognitive impairment patients.

BACKGROUND: Mild cognitive impairment (MCI) is a syndrome that disrupts an individual's cognitive function but preserves activities of daily living. MCI is thought to be a prodromal stage of dementia, which disrupts patients' daily lives and causes severe cognitive dysfunction. Although extensive clinical trials have attempted to slow or stop the MCI to dementia conversion, the results have been largely unsuccessful. The purpose of this study was to determine whether noninvasive electrical stimulation of MCI changes glucose metabolism. METHODS: Sixteen MCI patients participated in this study. We used transcranial direct current stimulation (tDCS) (2 mA/day, three times per week for 3 weeks) and assessed positron emission tomography (18 F-FDG) before and after 3 weeks of stimulation. RESULTS: We showed that regular and relatively long-term use of tDCS significantly increased regional cerebral metabolism in MCI patients. Furthermore, subjective memory satisfaction and improvement of the memory strategies of participants were observed only in the real tDCS group after 3 weeks of stimulation. CONCLUSION: Our findings suggest that neurophysiological intervention of MCI could improve glucose metabolism and transient memory function in MCI patients.

Reappraising abstract paintings after exposure to background information.

Can knowledge help viewers when they appreciate an artwork? Experts' judgments of the aesthetic value of a painting often differ from the estimates of naïve viewers, and this phenomenon is especially pronounced in the aesthetic judgment of abstract paintings. We compared the changes in aesthetic judgments of naïve viewers while they were progressively exposed to five pieces of background information. The participants were asked to report their aesthetic judgments of a given painting after each piece of information was presented. We found that commentaries by the artist and a critic significantly increased the subjective aesthetic ratings. Does knowledge enable experts to attend to the visual features in a painting and to link it to the evaluative conventions, thus potentially causing different aesthetic judgments? To investigate whether a specific pattern of attention is essential for the knowledge-based appreciation, we tracked the eye movements of subjects while viewing a painting with a commentary by the artist and with a commentary by a critic. We observed that critics' commentaries directed the viewers' attention to the visual components that were highly relevant to the presented commentary. However, attention to specific features of a painting was not necessary for increasing the subjective aesthetic judgment when the artists' commentary was presented. Our results suggest that at least two different cognitive mechanisms may be involved in knowledge- guided aesthetic judgments while viewers reappraise a painting.

Decoding covert motivations of free riding and cooperation from multi-feature pattern analysis of EEG signals.

Cooperation and free riding are among the most frequently observed behaviors in human social decision-making. In social interactions, the effects of strategic decision processes have been consistently reported in iterative cooperation decisions. However, the neural activity immediately after new information is presented, the time at which strategy learning potentially starts has not yet been investigated with high temporal resolution. Here, we implemented an iterative, binary public goods game that simulates cooperation/free riding behavior. We applied the multi-feature pattern analysis method by using a support vector machine and the unique combinatorial performance measure, and identified neural features from the single-trial, event-related spectral perturbation at the result-presentation of the current round that predict participants' decisions to cooperate or free ride in the subsequent round. We found that neural oscillations in centroparietal and temporal regions showed the highest predictive power through 10-fold cross-validation; these predicted the participants' next decisions, which were independent of the neural responses during their own preceding choices. We suggest that the spatial distribution and time-frequency information of the selected features represent covert motivations to free ride or cooperate in the next round and are separately processed in parallel with information regarding the preceding results.

Beta-frequency EEG activity increased during transcranial direct current stimulation.

Transcranial direct current stimulation (tDCS) is a technique for noninvasively stimulating specific cortical regions of the brain with small (<2 mA) and constant direct current on the scalp. tDCS has been widely applied, not only for medical treatment, but also for cognitive and somatosensory function enhancement, motor learning improvement, and social behavioral change. However, the mechanism that underlies the effect of tDCS is unclear. In this study, we performed simultaneous electroencephalogram (EEG) monitoring during tDCS to understand the dynamic electrophysiological changes throughout the stimulation. A total of 10 healthy individuals participated in this experiment. We recorded EEGs with direct current stimulation, as well as during a 5-min resting state before and after the stimulation. All participants kept their eyes closed during the experiment. Anode and cathode patches of tDCS were placed on the left and the right dorsolateral prefrontal cortex, respectively. In addition, an EEG electrode was placed on the medial prefrontal cortex. The beta-frequency power increased promptly after starting the stimulation. The significant beta-power increase was maintained during the stimulation. Other frequency bands did not show any significant changes. The results indicate that tDCS of the left dorsolateral prefrontal cortex changed the brain to a ready state for efficient cognitive functioning by increasing the beta-frequency power. This is the first attempt to simultaneously stimulate the cortex and record the EEG and then systematically analyze the prestimulation, during-stimulation, and poststimulation EEG data.

Alcohol reduces cross-frequency theta-phase gamma-amplitude coupling in resting electroencephalography.

BACKGROUND: The electrophysiological inhibitory mechanism of cognitive control for alcohol remains largely unknown. The purpose of the study was to compare electroencephalogram (EEG) power spectra and cross-frequency phase-amplitude coupling (CFPAC) at rest and during a simple subtraction task after acute alcohol ingestion. METHODS: Twenty-one healthy subjects participated in this study. Two experiments were performed 1 week apart, and the order of the experiments was randomly assigned to each subject. During the experiments, each subject was provided with orange juice containing alcohol or orange juice only. We recorded EEG activity and analyzed power spectra and CFPAC data. RESULTS: The results showed prominent theta-phase gamma-amplitude coupling at the frontal and parietal electrodes at rest. This effect was significantly reduced after alcohol ingestion. CONCLUSIONS: Our findings suggest that theta-phase gamma-amplitude coupling is deficiently synchronized at rest after alcohol ingestion. Therefore, cross-frequency coupling could be a useful tool for studying the effects of alcohol on the brain and investigating alcohol addiction.

Interpersonal body and neural synchronization as a marker of implicit social interaction.

One may have experienced his or her footsteps unconsciously synchronize with the footsteps of a friend while walking together, or heard an audience's clapping hands naturally synchronize into a steady rhythm. However, the mechanisms of body movement synchrony and the role of this phenomenon in implicit interpersonal interactions remain unclear. We aimed to evaluate unconscious body movement synchrony changes as an index of implicit interpersonal interaction between the participants, and also to assess the underlying neural correlates and functional connectivity among and within the brain regions. We found that synchrony of both fingertip movement and neural activity between the two participants increased after cooperative interaction. These results suggest that the increase of interpersonal body movement synchrony via interpersonal interaction can be a measurable basis of implicit social interaction. The paradigm provides a tool for identifying the behavioral and the neural correlates of implicit social interaction.

Modulation of theta phase synchronization in the human electroencephalogram during a recognition memory task.

To the extent that recognition memory relies on interactions among widely distributed neural assemblies across the brain, phase synchronization between brain rhythms may play an important role in meditating those interactions. As the theta rhythm is known to modulate in power during the recognition memory process, we aimed to determine how the phase synchronization of the theta rhythms across the brain changes with recognition memory. Fourteen human participants performed a visual object recognition task in a virtual reality environment. Electroencephalograms were recorded from the scalp of the participants while they either recognized objects that had been presented previously or identified new objects. From the electroencephalogram recordings, we analyzed the phase-locking value of the theta rhythms, which indicates the degree of phase synchronization. We found that the overall phase-locking value recorded during the recognition of previously viewed objects was greater than that recorded during the identification of new objects. Specifically, the theta rhythms became strongly synchronized between the frontal and the left parietal areas during the recognition of previously viewed objects. These results suggest that the recognition memory process may involve an interaction between the frontal and the left parietal cortical regions mediated by theta phase synchronization.

Decreased cortical complexity in methamphetamine abusers.

This study aimed to investigate if methamphetamine (MA) abusers exhibit alterations in complexity of the electroencephalogram (EEG) and to determine if these possible alterations are associated with their abuse patterns. EEGs were recorded from 48 former MA-dependent males and 20 age- and sex-matched healthy subjects. Approximate Entropy (ApEn), an information-theoretical measure of irregularity, of the EEGs was estimated to quantify the degree of cortical complexity. The ApEn values in MA abusers were significantly lower than those of healthy subjects in most of the cortical regions, indicating decreased cortical complexity of MA abusers, which may be associated with impairment in specialization and integration of cortical activities owing to MA abuse. Moreover, ApEn values exhibited significant correlations with the clinical factors including abuse patterns, symptoms of psychoses, and their concurrent drinking and smoking habits. These findings provide insights into abnormal information processing in MA abusers and suggest that ApEn of EEG recordings may be used as a potential supplementary tool for quantitative diagnosis of MA abuse. This is the first investigation to assess the "severity-dependent dynamical complexity" of EEG patterns in former MA abusers and their associations with the subjects' abuse patterns and other clinical measures.

Mathematically gifted adolescents have deficiencies in social valuation and mentalization.

Many mathematically gifted adolescents are characterized as being indolent, underachieving and unsuccessful despite their high cognitive ability. This is often due to difficulties with social and emotional development. However, research on social and emotional interactions in gifted adolescents has been limited. The purpose of this study was to observe differences in complex social strategic behaviors between gifted and average adolescents of the same age using the repeated Ultimatum Game. Twenty-two gifted adolescents and 24 average adolescents participated in the Ultimatum Game. Two adolescents participate in the game, one as a proposer and the other as a responder. Because of its simplicity, the Ultimatum Game is an apt tool for investigating complex human emotional and cognitive decision-making in an empirical setting. We observed strategic but socially impaired offers from gifted proposers and lower acceptance rates from gifted responders, resulting in lower total earnings in the Ultimatum Game. Thus, our results indicate that mathematically gifted adolescents have deficiencies in social valuation and mentalization.