Machine learning and big data analytics in bipolar disorder: A position paper from the International Society for Bipolar Disorders Big Data Task Force.

OBJECTIVES: The International Society for Bipolar Disorders Big Data Task Force assembled leading researchers in the field of bipolar disorder (BD), machine learning, and big data with extensive experience to evaluate the rationale of machine learning and big data analytics strategies for BD. METHOD: A task force was convened to examine and integrate findings from the scientific literature related to machine learning and big data based studies to clarify terminology and to describe challenges and potential applications in the field of BD. We also systematically searched PubMed, Embase, and Web of Science for articles published up to January 2019 that used machine learning in BD. RESULTS: The results suggested that big data analytics has the potential to provide risk calculators to aid in treatment decisions and predict clinical prognosis, including suicidality, for individual patients. This approach can advance diagnosis by enabling discovery of more relevant data-driven phenotypes, as well as by predicting transition to the disorder in high-risk unaffected subjects. We also discuss the most frequent challenges that big data analytics applications can face, such as heterogeneity, lack of external validation and replication of some studies, cost and non-stationary distribution of the data, and lack of appropriate funding. CONCLUSION: Machine learning-based studies, including atheoretical data-driven big data approaches, provide an opportunity to more accurately detect those who are at risk, parse-relevant phenotypes as well as inform treatment selection and prognosis. However, several methodological challenges need to be addressed in order to translate research findings to clinical settings.

Relationship among latitude, climate, season and self-reported mood in bipolar disorder.

OBJECTIVE: Many researchers have analyzed seasonal variation in hospital admissions for bipolar disorder with inconsistent results. We investigated if a seasonal pattern was present in daily self-reported daily mood ratings from patients living in five climate zones in the northern and southern hemispheres. We also investigated the influence of latitude and seasonal climate variables on mood. METHOD: 360 patients who were receiving treatment as usual recorded mood daily (59,422 total days of data). Both the percentage of days depressed and hypomanic/manic, and the episodes of depression and mania were determined. The observations were provided by patients from different geographic locations in North and South America, Europe and Australia. These data were analyzed for seasonality by climate zone using both a sinusoidal regression and the Gini index. Additionally, the influence of latitude and climate variables on mood was estimated using generalized linear models for each season and month. RESULTS: No seasonality was found in any climate zone by either method. In spite of vastly different weather, neither latitude nor climate variables were associated with mood by season or month. CONCLUSION: Daily self-reported mood ratings of most patients with bipolar disorder did not show a seasonal pattern. Neither climate nor latitude has a primary influence on the daily mood changes of most patients receiving medication for bipolar disorder.

Genetic studies of bipolar disorder in patients selected by their treatment response

Bipolar disorder (BD) is a major mood disorder with several genes of moderate or small effect contributing to the genetic susceptibility. It is also likely heterogeneous, which stimulated efforts to refine its clinical phenotype, studies investigating the link between BD susceptibility and response to a specific mood stabilizer appear to be one of the promising directions. In particular, excellent response to lithium prophylaxis has been described as a clinical marker of a more homogeneous subgroup of BD, characterized by an episodic course, low rates of co-morbid conditions, absence of rapid cycling, and a strong genetic loading. These results also suggest that lithium response clusters in families (independent of the increased familial loading for affective disorders), likely on a genetic basis. For almost 40 years, clinical studies have pointed to differences between lithium responders (LR) and non-responders (LNR). For instance, there is a higher frequency of BD in LR families. As well, investigations in offspring of LR and LNR probands show that the offspring of LR tend to manifest a higher frequency of affective disorders, less co-morbidity and an episodic course of the disorder, compared with the offspring of LNR, who had a broad range of psychopathology, a higher rate of co-morbidity and a chronic course of the disorder. A number of candidate genes have been studied in patients treated with lithium; of these, several showed an association in at least one study: cAMP responsive element binding protein (CREB), X-box binding protein 1 (XBP-1), inositol polyphosphate-1-phosphatase (INNP1), serotonin transporter gene (5-HTT), brain-derived neurotrophic factor (BDNF), phospholipase γ-1 (PLCγ-1), dopamine receptors (D2 and D4), polyglutamine tracts, tyrosine hydroxylase, inositol monophosphatase (IMPA), mitochondrial DNA, and breakpoint cluster region (BCR) gene. Clinical studies have shown as well that the treatment response and outcome appear to be specific for the different types of mood stabilizers. Patients who respond to lithium exhibit qualitative differences from patients responding to other medications, such as valproate, carbamazepine or lamotrigine. Responders to carbamazepine had atypical clinical features, such as mood-incongruent psychosis, an age at onset of illness below 30 years old, and a negative family history of mood disorders. Similarly, in a study comparing the phenotypic spectra in responders to lithium versus lamotrigine the probands differed with respect to clinical course (with rapid cycling and non-episodic course in the lamotrigine group) and co-morbidity, with the lamotrigine-responder group showing a higher frequency of panic attacks and substance abuse. In conclusion, pharmacogenetic studies may provide important clues to the nature of bipolar disorder and the response to long term treatment.

El trastorno bipolar (TB) es un trastorno afectivo con varios genes, de efecto leve o moderado, que contribuyen a su susceptibilidad. Es asimismo un trastorno heterogéneo, lo que ha estimulado diversas iniciativas para refinar el fenotipo de los pacientes con este trastorno. Particularmente en el TB, una respuesta excelente a la profilaxis con litio ha sido descrita como un marcador clínico en un subgrupo más homogéneo en TB, caracterizado por un curso episódico, baja prevalencia respecto a comorbilidad, ausencia de ciclado rápido y una carga genética importante. En relación con ello, y a pesar de que la totalidad de los estudios no coinciden, la mayor parte sugiere que seleccionar <> de acuerdo con su respuesta al tratamiento incrementa la homogeneidad fenotípica. Estos resultados sugieren asimismo que la respuesta al litio <> en familias (independientemente de la tasa familiar incrementada para trastornos afectivos), muy probablemente con bases genéticas. Por casi 40 años, los estudios clínicos han dilucidado las diferencias entre los respondedores a litio (LR) y los no respondedores (LNR). A este respecto, existe una frecuencia más alta de TB en familias LR; asimismo, las investigaciones en los descendientes de los probandos LR y LNR han demostrado que los descendientes de LR tienden a manifestar una mayor frecuencia de trastornos afectivos, menor comorbilidad, y un curso episódico del trastorno comparados con los descendientes de LNR, quienes muestran un amplio espectro de psicopatología, una alta tasa de comorbilidad, y un curso crónico del trastorno. Diversos genes candidatos han sido estudiados en pacientes tratados con litio, y varios de ellos han mostrado una asociación en al menos un estudio: proteína de unión al elemento de respuesta (cAMP responsive element binding protein, CREB), proteína de unión a X-box 1 (X-box binding protein 1, XBP-1), inositol polifosfato-1-fosfatasa (INNP1), transportador de serotonina (5-HTT), factor de crecimiento derivado del cerebro (brain-derived neurotrophic factor, BDNF), fosfolipasa γ-1(PLCγ-1), receptores dopaminérgicos (D2 y D4), poliglutamina, tirosina hidroxilasa, inositol monofosfatasa (IMPA), DNA mitocondrial y el gen BCR. Los estudios clínicos han demostrado que la respuesta al tratamiento y el pronóstico parecen ser específicos para los diferentes tipos de estabilizadores del ánimo. Los pacientes que responden al litio exhiben diferencias cualitativas respecto a los pacientes que responden a otros estabilizadores del ánimo, como valproato, carbamazepina o lamotrigina. Los respondedores a carbamazepina presentan características clínicas atípicas, como psicosis incongruente con el afecto, una edad de inicio del trastorno menor a 30 años de edad y no cuentan con historia familiar de trastornos afectivos. Finalmente, un estudio entre los respondedores al litio y los respondedores a lamotrigina demostró que los probandos difieren con respecto al curso clínico del trastorno (con ciclado rápido y un curso no episódico en los pacientes que responden a lamotrigina), así como en la comorbilidad, teniendo los pacientes que responden a lamotrigina una mayor frecuencia de crisis de angustia y abuso de sustancias. En conclusión, los estudios farmacogenéticos podrían proveer de hallazgos importantes respecto a la naturaleza del trastorno bipolar y la respuesta al tratamiento a largo plazo.

Transtorno bipolar / Bipolar disorder

Os resultados de estudos de famílias sugerem que o transtorno bipolar tenha uma base genética. Essa hipótese foi reforçada em estudos de adoção e de gêmeos. A herança do transtorno bipolar é complexa, envolve vários genes, além de apresentar heterogeneidade e interação entre os fatores genéticos e não-genéticos. Achados, que já foram replicados, já implicaram os cromossomos 4, 12, 18 e 21, entre outros, na busca por genes de suscetibilidade. Os resultados mais promissores foram obtidos através de estudos de ligação. Por outro lado, os estudos de associação geraram dados interessantes, mas ainda vagos. Os estudos de populações de pacientes homogêneos e a melhor definição do fenótipo deverão contribuir para avanços futuros. A identificação dos genes relacionados ao transtorno bipolar irá permitir o melhor entendimento e tratamento dessa doença