Machine Learning Enhances the Efficiency of Cognitive Screenings for Primary Care.

BACKGROUND: Incorporation of cognitive screening into the busy primary care will require the development of highly efficient screening tools. We report the convergence validity of a very brief, self-administered, computerized assessment protocol against one of the most extensively used, clinician-administered instruments-the Montreal Cognitive Assessment (MoCA). METHOD: Two hundred six participants (mean age = 67.44, standard deviation [SD] = 11.63) completed the MoCA and the computerized test. Three machine learning algorithms (ie, Support Vector Machine, Random Forest, and Gradient Boosting Trees) were trained to classify participants according to the clinical cutoff score of the MoCA (ie, < 26) from participant performance on 25 features of the computerized test. Analysis employed Synthetic Minority Oversampling TEchnic to correct the sample for class imbalance. RESULTS: Gradient Boosting Trees achieved the highest performance (accuracy = 0.81, specificity = 0.88, sensitivity = 0.74, F1 score = 0.79, and area under the curve = 0.81). A subsequent K-means clustering of the prediction features yielded 3 categories that corresponded to the unimpaired (mean = 26.98, SD = 2.35), mildly impaired (mean = 23.58, SD = 3.19), and moderately impaired (mean = 17.24, SD = 4.23) ranges of MoCA score ( F = 222.36, P < .00). In addition, compared to the MoCA, the computerized test correlated more strongly with age in unimpaired participants (ie, MoCA ≥26, n = 165), suggesting greater sensitivity to age-related changes in cognitive functioning. CONCLUSION: Future studies should examine ways to improve the sensitivity of the computerized test by expanding the cognitive domains it measures without compromising its efficiency.

Stigma, social anxiety, and illness severity in bipolar disorder: Implications for treatment.

BACKGROUND: Studies indicate that comorbid anxiety disorders predict a more severe course of illness in bipolar disorder (BD). The relatively high prevalence of social anxiety in BD points to the potential role that socio-cultural factors, such as stigma, play in exacerbating the progression of this disorder. Stigma creates social anxiety in affected individuals because it essentially forces them into a vulnerable social status that is marked by public disgrace. Although the etiology of debilitating social anxiety in BD may involve multiple factors, stigma deserves particular clinical attention because research in this area indicates that it is common and its internalization is associated with poor outcome. METHODS: We conducted a literature review using search terms related to stigma, social anxiety, bipolar disorder, illness severity, and outcomes. The electronic databases searched included PsychINFO, PubMed, JSTOR, and EBSCOhost Academic Search Complete with limits set to include articles published in English. RESULTS: The literature indicates that internalized stigma often triggers the core psychological experiences of social anxiety and is highly correlated with clinical and functional outcome in BD. On a psychological level, internalized stigma and social anxiety can create distress that triggers symptoms of BD. From a biological perspective, stigma constitutes a chronic psychosocial stressor that may interact with the pathophysiology of BD in inflammatory ways. CONCLUSIONS: The connection between stigma and social anxiety, and their combined effects on people with BD, carries important implications for psychiatric care. To obtain an accurate clinical formulation, initial evaluations may seek to examine stigma-related experiences and determine their relationship to anxiety symptoms and psychosocial functioning. In addition, direct interventions for reducing the ill effects of stigma in BD deserve clinical attention, because they may carry the potential to enhance outcomes.

Autonomic nervous system arousal and cognitive functioning in bipolar disorder.

OBJECTIVE: Previous theories about the etiology of cognitive dysfunction in bipolar disorder (BD) emphasized trait factors such as neurological impairment. State factors, other than mood symptoms, that may exacerbate functional deficits have not yet been considered. The purpose of this study was to examine autonomic nervous system (ANS) arousal following cognitive challenge. The study compared patients with BD and healthy controls (HC) in physiological measures and neuropsychological test scores. METHODS: Thirty euthymic patients with BD and 22 HC completed the study. Participants completed mood [Beck Depression Inventory-II (BDI-II) and Young Mania Rating Scale (YMRS)], anxiety (State-Trait Anxiety Inventory), and substance abuse (Drug Abuse Screening Test-20 item and Alcohol Use Disorders Identification Test) measures. They were connected to an electrogram, a sensitive thermometer for measuring finger temperature, and electrodes that measure galvanic skin response. After a five-min baseline measurement in a restful state, participants completed a computerized neuropsychological battery (CNS Vital Signs). RESULTS: The group with BD reported significantly more mood symptoms (BDI-II, t = 3.71, p < 0.001; YMRS, t = 6.73, p < 0.001) and scored higher on a measure of trait-anxiety (State-Trait Anxiety Inventory, t = 2.91, p < 0.001) than HC. A multivariate analysis of variance revealed higher arousal on all physiological measures in the BD group relative to HC at baseline [F(3,48) = 13.1, p < 0.001] and during cognitive testing [F(3,48) = 11.3, p < 0.001]. The increase in physiological arousal from a restful state to the time of testing was higher for the BD group [F(3,37) = 8.06, p < 0.001]. With respect to cognitive data, HC scored higher than patients with BD across the measures of memory (F = 8.5, p < 0.001), sustained (F = 9.5, p < 0.001) and complex (F = 2.7, p < 0.04) attention, processing speed (F = 10.0, p < 0.001), reaction time (F = 7.8, p < 0.001), cognitive flexibility (F = 19.7, p < 0.001), working memory (F = 10.8, p < 0.001), and social acuity (F = 5.7, p < 0.01), with partial eta-squared from 0.18 to 0.62. Correlational analysis revealed significant associations between various cognitive test scores and changes in physiological arousal from baseline to testing (-0.59 ≤ r ≤ 0.22). CONCLUSIONS: Relative to HC, patients with BD experience larger changes in ANS arousal between a restful baseline and cognitive testing, and achieve lower cognitive test scores. Further research is needed to determine whether acute physiological symptoms of anxiety directly compromise cognitive functioning in BD.

Cognitive and psychosocial functioning in bipolar disorder with and without psychosis during early remission from an acute mood episode: a comparative longitudinal study.

BACKGROUND: The current investigation aimed to extend previous findings, which linked psychosis in bipolar disorder (BD) to cognitive impairment during hospital discharge and readmission, by examining the recovery of patients with psychosis who were not re-hospitalized. The study compared mood, cognitive and functional outcomes in patients who had, versus had not, experienced psychosis during a recent psychiatric hospitalization. The hypothesis was that patients admitted to the hospital with psychosis would exhibit more residual symptoms, greater cognitive deficits, and lower psychosocial functioning than patients who presented to care without psychosis. Group differences were expected to emerge both at the time of hospital discharge and at a 3-month follow up. METHOD: Fifty-five participants (ages 18-59, 25 women, 20 with psychosis) with BD I disorder completed both assessments, which included a clinical and diagnostic interview, functional evaluation, and the administration of mood measures and a neuropsychological battery. RESULTS: The groups were comparable with respect to illness history (e.g., number of previous hospitalizations, age of onset, employment). At discharge and follow-up, the group with psychosis exhibited more mood symptoms, obtained lower GAF scores, and performed more poorly on measures of memory and executive functioning. At follow-up, participants with psychosis exhibited poorer psychosocial adaptation. LIMITATIONS: It is possible that some of the observed group differences in cognitive functioning emerged due to differences in medication efficacy or side effects. CONCLUSION: The results of this study support the hypothesis that psychosis in BD predicts limited recovery during early remission from mood disturbance, regardless of illness history.

Recovery of cognitive functioning in patients with co-occurring bipolar disorder and alcohol dependence during early remission from an acute mood episode.

OBJECTIVE: This study aimed to examine cognitive recovery in patients with co-occurring bipolar disorder (BD) and alcohol dependence (AD) during remission from an acute mood disturbance. METHOD: Fifty-five adult inpatients with bipolar I disorder (BD I) completed a neuropsychological battery, mood measures, and substance abuse measures upon discharge from the hospital and at a 3-month follow-up. Analyses provided group comparisons on these measures between patients who presented with co-occurring AD (n = 21) in the year prior to hospital admission and patients without a substance use disorder (SUD; n = 34). RESULTS: Multivariate analyses of variance detected group differences on measures of visual memory, verbal memory, and executive functioning, using previous number of psychiatric admissions and age of onset of BD as covariates. These differences occurred both at discharge and follow-up. Between discharge and follow-up, the group without SUD exhibited more substantial gains than the group of dually diagnosed patients on free recall of verbal and visual materials and on a measure of cognitive flexibility. CONCLUSIONS: Patients with co-occurring BD and AD may suffer from more severe cognitive dysfunction and less favorable recovery of cognitive deficits than patients without SUD over the course of remission from a mood episode.

Toward Pre-Diagnostic Detection of Dementia in Primary Care.

BACKGROUND: Preventing dementia warrants the pragmatic engagement of primary care. OBJECTIVE: This study predicted conversion to dementia 12 months before diagnosis with indicators that primary care can utilize within the practical constraints of routine practice. METHODS: The study analyzed data from the Alzheimer's Disease Neuroimaging Initiative (Total sample = 645, converting participants = 54). It predicted the conversion from biological (plasma neurofilament light chain), cognitive (Trails Making Test- B), and functional (Functional Activities Questionnaire) measures, in addition to demographic variables (age and education). RESULTS: A Gradient Booster Trees classifier effectively predicted the conversion, based on a Synthetic Minority Oversampling Technique (n = 1,290, F1 Score = 92, AUC = 94, Recall = 87, Precision = 97, Accuracy = 92). Subsequent analysis indicated that the MCI False Positive group (i.e., non-converting participants with cognitive impairment flagged by the model for prospective conversion) scored significantly lower on multiple cognitive tests (Montreal Cognitive Assessment, p < 0.002; ADAS-13, p < 0.0004; Rey Auditory Verbal Learning Test, p < 0.002/0.003) than the MCI True Negative group (i.e., correctly classified non-converting participants with cognitive impairment). These groups also differed in CSF tau levels (p < 0.04), while consistent effect size differences emerged in the all-pairwise comparisons of hippocampal volume and CSF Aß1 - 42. CONCLUSION: The model effectively predicted 12-month conversion to dementia and further identified non-converting participants with MCI, in the False Positive group, at relatively higher neurocognitive risk. Future studies may seek to extend these results to earlier prodromal phases. Detection of dementia before diagnosis may be feasible and practical in primary care settings, pending replication of these findings in diverse clinical samples.

Ecologically valid support for the link between cognitive and psychosocial functioning in bipolar disorder.

Prior research into the link between cognitive and psychosocial functioning in bipolar disorder has examined primarily asymptomatic patients, has measured these domains concurrently, and has failed to establish convergent validity in the assessment of psychosocial dysfunction. The present study examines the relation between cognitive and psychosocial functioning at the time of discharge from hospitalization for acute mood disturbance. We obtained measures of psychosocial functioning that were both close and distant to the time of neuropsychological testing; the former from the discharging psychiatrists, and the latter from reports of formally recognized disability status, determined by persons wholly unrelated to the present research. Sixty-three patients with bipolar I disorder, hospitalized for acute mood disturbance, completed a neuropsychological test battery 24 to 48 h prior to discharge. We compared patients with versus without formal disability status on the Global Assessment of Functioning (GAF) scale and on scores of neuropsychological tests. We also tested associations between GAF scores and cognitive test scores. Results supported the convergent validity in the measurement of psychosocial disability, underscored the robust connection between cognitive and psychosocial impairment, and highlighted the presence of this connection during an important clinical state - time of discharge from psychiatric hospitalization.

Subjective Versus Objective Assessment of Cognitive Functioning in Primary Care.

PURPOSE: This study examined the clinical utility of highly efficient subjective and objective screens of cognitive impairment. METHOD: Participants (N = 124, age ≥ 65, mean = 73.59, SD = 6.26) completed a 2-item questionnaire of subjective memory functioning, a brief computerized cognitive test, and the Montreal Cognitive Assessment (MoCA). Next, participants were assigned to 1 of 4 conditions, based on their subjective (low/high) and objective (impaired/unimpaired) levels of cognitive functioning. Further analysis divided the sample into age-based groups (ie, age < 75, age ≥ 75). RESULTS: The proportion of participants in the impaired subsample (ie, MoCA < 26), who reported a high level of subjective concern about their memory, was low (ie, 0.15). Among unimpaired participants, analysis detected significant group differences across subjective memory levels (P < .0003) and age (P < .005) categories on one of the three tasks of the computerized test (ie, cognitive control). In contrast, the MoCA offered no differentiation between these groups. CONCLUSION: Screening protocols in which cognitive testing is administered subsequent to patient complaint are prone to underdiagnosis. In addition, common dementia screens are insensitive to subjective deficits and healthy cognitive aging. Therefore, they may lead to dismissing valid concerns that deserve preventive attention. Primary care needs efficient screening tools that are sensitive to prodromal decline.

The duration of inpatient admission predicts cognitive functioning at discharge in patients with bipolar disorder.

Previous studies of cognitive functioning in bipolar disorder generally indicate that a more severe course of illness is associated with greater cognitive impairment. In particular, a history of greater number and longer duration of mood episodes predicts enduring cognitive deficits in euthymic patients. Shifting the focus of this investigation to the cognitive effects of a discrete mood episode, the current study aimed to explore whether patients who require a longer hospitalization to stabilize from an acute episode of mood disturbance present with more compromised cognitive functioning during the phase of early recovery. For this purpose, the study examined the link between the duration of inpatient admission and neuropsychological test scores at the time of discharge in 41 patients with bipolar disorder. Participants were assigned to long (n = 20) and short (n = 21) stay groups using a median split (M = 12). Results indicated that longer admissions were associated with more severe deficits in executive functioning at discharge after controlling for residual mood symptoms and previous number of psychiatric admissions. Findings from the current study may inform discharge planning for patients with bipolar disorder after an extended hospital stay.

Neurocognitive impairment in patients with co-occurring bipolar disorder and alcohol dependence upon discharge from inpatient care.

The current study explored the neurocognitive functioning of patients with co-occurring bipolar disorder and alcohol dependence upon discharge from inpatient care. The study compared scores of neuropsychological tests among three groups of bipolar I inpatients without a history of neurological injury or illness: 1) patients meeting DSM-IV diagnostic criteria for alcohol dependence in the past 6 months (n=13), 2) patients diagnosed with alcohol dependence in full remission (n=9), and 3) patients without a history of a substance use disorder (SUD; n=41). Analyses indicated that patients with co-occurring alcohol dependence exhibited more severe impairment on tests of executive functioning (i.e. Stroop Color-Word Interference Test, Wisconsin Card Sorting Test) than patients without SUD. In addition, the group meeting diagnostic criteria for alcohol dependence in the past 6 months exhibited greater decrements in verbal (California Verbal Learning Test--II) and visual (Rey Complex Figure Test) memory. Analysis further indicated that patients in full SUD remission scored lower on measures of fluid intelligence (Wechsler Abbreviated Scale of Intelligence--Performance IQ). Consistent with previous reports, in the current sample, co-occurring alcohol dependence predicted higher rates of disability status. It is possible that cognitive deficits of greater severity in dually diagnosed patients contribute to this unfavorable outcome. Recognizing the extent of cognitive impairment in dually diagnosed patients may facilitate the effort to ameliorate their condition.

Increasing the power for detecting impairment in older adults with the Faces subtest from Wechsler Memory Scale-III: an empirical trial.

Empirical studies have questioned the validity of the Faces subtest from the WMS-III for detecting impairment in visual memory, particularly among the elderly. A recent examination of the test norms revealed a significant age related floor effect already emerging on Faces I (immediate recall), implying excessive difficulty in the acquisition phase among unimpaired older adults. The current study compared the concurrent validity of the Faces subtest with an alternative measure between 16 Alzheimer's patients and 16 controls. The alternative measure was designed to facilitate acquisition by reducing the sequence of item presentation. Other changes aimed at increasing the retrieval challenge, decreasing error due to guessing and standardizing the administration. Analyses converged to indicate that the alternative measure provided a considerably greater differentiation than the Faces subtest between Alzheimer's patients and controls. Steps for revising the Faces subtest are discussed.

Developing Cognitive Markers of Alzheimer's Disease for Primary Care: Implications for Behavioral and Global Prevention.

A comprehensive approach to the prevention of Alzheimer's disease (AD) warrants a synergy across multiple domains and procedures. Whereas the study of biological markers has mobilized major activity in the field, the development of cognitive markers is largely ignored, despite the unique advantages they may offer. Cognitive markers essentially assess the core clinical feature that biological markers intend to predict. In this respect, cognitive markers expand the foundation of preclinical diagnostics and disease staging in a manner that integrates both physiological and psychological factors. In addition, the cost-effective implementation of cognitive markers makes them remarkably conducive to community-wide screenings, and thereby a vital component of any global blueprint for prevention. Specifically, in the primary care setting, cognitive markers may provide effective gate keeping for more invasive, labor intensive, and expensive procedures. From this perspective, cognitive markers may provide the first step for identifying preclinical treatment recipients in general public. Moreover, the detection of preclinical decline via cognitive markers can increase awareness of AD risk and the motivation for making protective lifestyle changes. The behavioral approach might be expedient for prevention in light of the compelling evidence of lifestyle amelioration of AD risk. In an integrative view, incorporating cognitive markers to primary care may facilitate a synergetic development in preventive interventions that carries epidemiological significance. This paper addresses the theoretical, methodological, and pragmatic aspects of this prospect.

Reinforcement expectancies for studying predict drinking problems among college students: approaching problem drinking from an expectancies choice perspective.

Molar behavioral choice theories point to the important role reinforcement for alternative behaviors may play in reducing problem drinking [Psychol. Addict. Behav. 9 (1995) 223.]. A similar molar choice perspective may apply to a cognitive mediational level of analysis: the molecular link between alcohol expectancies and drinking behavior depends on expectancies for alternative behaviors. In this study, 82 college students completed the Alcohol Effects Questionnaire (AEFQ), the Rutgers Alcohol Problem Index (RAPI), a self-report measure of drinking quantity and frequency, and a five-item studying expectancies questionnaire devised for the current study. Analyses indicated that, among students who hold high alcohol expectancies, those who reported high studying expectancies drink less and develop significantly fewer drinking problems than those who hold low studying expectancies. Future research may address the possibility of reducing drinking problems among college students by enhancing studying expectancies.

Discriminating reinforcement expectancies for studying from future time perspective in the prediction of drinking problems.

In recent work, Levy and Earleywine [Addict. Behav. (2003)] have proposed an expectancy framework of choice for the study of alcohol abuse. Within this framework, the link between alcohol expectancies and drinking problems is studied in the context of expectancies for alternative behaviors. Preliminary findings showed that studying expectancies predicted drinking problems among college students who hold high positive alcohol expectancies. Since the expected gain from studying is inherently long-term, further validation of these results requires to discriminate studying expectancies from the personality construct of future time perspective (FTP), which has been previously shown to predict substance use. In this study, studying expectancies produced larger group differences than FTP on all measures of drinking problems and habits. These differences remained significant after controlling for FTP. The results of the current study thus replicate previous findings and provide further support for an expectancy model of choice. Future research may explore whether enhancing studying expectancies can reduce drinking problems among college and high school students.

Illness severity, trait anxiety, cognitive impairment and heart rate variability in bipolar disorder.

Numerous studies have documented a significant association between symptom severity and cognitive functioning in bipolar disorder (BD). These findings advanced speculations about a potential link between the physiological stress associated with illness severity and cognitive dysfunction. To explore this hypothesis, the current study employed heart rate variability (HRV) as a physiological measure that is sensitive to the effects of chronic stress, and a scale of trait anxiety for assessing a psychological condition that is correlated with hyper sympathetic arousal. Analyses indicated that BD patients with High Illness Severity reported more symptoms of trait-anxiety (i.e., State Trait Anxiety Inventory), performed more poorly on a computerized neuropsychological battery (i.e., CNS Vital Signs), and exhibited a more constricted HRV profile (i.e., lower SDNN with elevated LF/HF ratio) than patients with Low Illness Severity. Illness severity was determined by a history of psychosis, illness duration, and number of mood episodes. A third group of healthy controls (n=22) performed better on the neuropsychological battery and exhibited a healthier HRV profile than the BD groups. This study provides preliminary evidence that illness severity and cognitive impairment in BD may be associated with state anxiety and neuro-cardiac alterations that are sensitive to physiological stress.

Neuropsychological screening as a standard of care during discharge from psychiatric hospitalization: the preliminary psychometrics of the CNS Screen.

Cost-prohibitive factors currently prevent a warranted integration of neuropsychological screenings into routine psychiatric evaluations, as a standard of care. To overcome this challenge, the current study examined the psychometric properties of a new computerized measure-the CNS Screen. One hundred and twenty six psychiatric inpatients completed the CNS Screen, the Montreal Cognitive Assessment (MoCA), and the Quick Inventory of Depressive Symptomatology-Self Rated (QIDS-SR16) on the day of hospital discharge. Statistical analysis established convergent validity with a moderate correlation between the self-administered CNS Screen and the clinician-administered MoCA (r=0.64). Discriminant validity was implicated by a non-significant correlation with the QIDS-SR16. Concurrent validity was supported by a moderate, negative correlation with patients' age (r=-0.62). In addition, consistent with previous findings, patients with psychotic disorders exhibited significantly poorer performance on the CNS Screen than patients with a mood disorder. Similarly, patients with a formal disability status scored significantly lower than other patients. The CNS Screen was well tolerated by all patients. With further development, this type of measure may provide a cost-effective approach to expanding neuropsychological screenings on inpatient psychiatric units.

Functional outcome in bipolar disorder: the big picture.

Previous research on functional outcome in bipolar disorder (BD) has uncovered various factors that exacerbate psychosocial disability over the course of illness, including genetics, illness severity, stress, anxiety, and cognitive impairment. This paper presents an integrated view of these findings that accounts for the precipitous decline in psychosocial functioning after illness onset. The proposed model highlights a number of reciprocal pathways among previously studied factors that trap people in a powerful cycle of ailing forces. The paper discusses implications to patient care as well as the larger social changes required for shifting the functional trajectory of people with BD from psychosocial decline to growth.

The characteristics of a discrete mood episode, neuro-cognitive impairment and re-hospitalization in bipolar disorder.

This longitudinal study examined characteristics of a discrete mood episode that predict re-hospitalization for bipolar disorder, highlighting associated cognitive dysfunction as a potential mechanism linking episode severity and relapse. Eighty-two inpatients meeting DSM-IV-TR diagnostic criteria for bipolar I disorder completed the study. Twenty-two of the patients were readmitted to the hospital within 3 months of discharge. The study compared these patients to the remaining 60 patients who were not readmitted to the hospital during this period. Patients were compared on several factors related to the severity of the mood episode and the course of illness more generally. Analysis also compared the groups on measures of mood and neuro-cognitive functioning, assessed 24-48 h before initial hospitalization discharge. Re-hospitalized patients had longer initial hospital stays (t = -3.10, p < 0.01), higher rates of psychosis while in the hospital (Chi square = 5.1, p < 0.02), and lower GAF scores on discharge (t = 2.37, p < 0.05). The groups did not differ in age of illness onset or number of previous psychiatric hospitalizations. With respect to neuro-cognitive functioning, analysis indicated poorer performance for re-hospitalized patients on measures of executive functioning (Wilks' Lambda, F (7, 71) = 9.0, p < 0.001), IQ (Wilks' Lambda, F (2, 76) = 5.06, p < 0.01), and memory (Wilks' Lambda, F (6,72) = 4.19, p < 0.001). Trends in the expected direction emerged for attention/working memory tests (Wilks' Lambda, F (7, 71) = 1.79, p < 0.10). Results highlight features of a discrete mood episode associated with increased rates of re-hospitalization. This study observed connections among episode severity, cognitive dysfunction at hospital discharge and re-hospitalization.

Cognitive impairment in bipolar disorder: an overview.

Research on cognitive impairment in bipolar disorder (BD) has prompted significant insights about the illness. New studies challenge previous notions about the episodic nature of BD, and account for psychosocial disability unrelated to mood disturbance. This article provides a conceptual overview of the growing body of research on cognitive dysfunction in BD. We discuss the evidence in light of the complexity inherent in the connection between cognitive deficits and neurological abnormalities in BD. This article also addresses issues related to etiology, advancing an integration of neurological, clinical, cognitive, and psychosocial factors into a model that elucidates how these factors interact to negatively impact persons with BD with a more severe course of illness. Lastly, this article discusses implications for patient care and future considerations.