Combining machine learning algorithms for prediction of antidepressant treatment response.

OBJECTIVES: Predictors for unfavorable treatment outcome in major depressive disorder (MDD) applicable for treatment selection are still lacking. The database of a longitudinal multicenter study on 1079 acutely depressed patients, performed by the German research network on depression (GRND), allows supervised and unsupervised learning to further elucidate the interplay of clinical and psycho-sociodemographic variables and their predictive impact on treatment outcome phenotypes. EXPERIMENTAL PROCEDURES: Treatment response was defined by a change of HAM-D 17-item baseline score ≥50% and remission by the established threshold of ≤7, respectively, after up to eight weeks of inpatient treatment. After hierarchical symptom clustering and stratification by treatment subtypes (serotonin reuptake inhibitors, tricyclic antidepressants, antipsychotic, and lithium augmentation), prediction models for different outcome phenotypes were computed with random forest in a cross-center validation design. In total, 88 predictors were implemented. RESULTS: Clustering revealed four distinct HAM-D subscores related to emotional, anxious, sleep, and appetite symptoms, respectively. After feature selection, classification models reached moderate to high accuracies up to 0.85. Highest accuracies were observed for the SSRI and TCA subgroups and for sleep and appetite symptoms, while anxious symptoms showed poor predictability. CONCLUSION: Our results support a decisive role for machine learning in the management of antidepressant treatment. Treatment- and symptom-specific algorithms may increase accuracies by reducing heterogeneity. Especially, predictors related to duration of illness, baseline depression severity, anxiety and somatic symptoms, and personality traits moderate treatment success. However, prospectives application of machine learning models will be necessary to prove their value for the clinic.

Structural MRI correlates for vulnerability and resilience to major depressive disorder.

BACKGROUND: In major depressive disorder (MDD), it is unclear to what extent structural brain changes are associated with depressive episodes or represent part of the mechanism by which the risk for illness is mediated. The aim of this study was to investigate whether structural abnormalities are related to risk for the development of MDD. METHODS: We compared healthy controls with a positive family history for MDD (HC-FHP), healthy controls with no family history of any psychiatric disease (HC-FHN) and patients with MDD. Groups were age- and sex-matched. We analyzed data from high-resolution magnetic resonance imaging using voxel-based morphometry. We performed small volume corrections for our regions of interest (hippocampus, dorsolateral [DLPFC] and dorsomedial prefrontal cortex [DMPFC], anterior cingulate cortex [ACC] and basal ganglia) using a family-wise error correction (p < 0.05) to control for multiple comparisons. RESULTS: There were 30 participants in the HC-FHP group, 64 in the HC-FHN group and 33 patients with MDD. The HC-FHP group had smaller right hippocampal and DLPFC grey matter volumes compared with the HC-FHN group, and even smaller right hippocampal volumes compared with patients with MDD. In addition, the HC-FHP group exhibited smaller white matter volumes in the DLPFC and left putamen but also greater volumes in 2 areas of the DMPFC compared with the HC-FHN group. Patients with MDD exhibited smaller volumes in the ACC, DMPFC, DLPFC and the basal ganglia compared with healthy controls. LIMITATIONS: The retrospective identification of family history might result in a bias toward unidentified participants in the control group at risk for MDD, diminishing the effect size. CONCLUSION: Volume reductions in the hippocampus and DLPFC might be associated with a greater risk for MDD. The HC-FHP group had smaller hippocampal volumes compared with patients with MDD, which is suggestive for neuroplastic effects of treatment. The HC-FHP group had not yet experienced a depressive episode and therefore might have been resilient and might have had some protective strategies. Whether resilience is associated with the larger white matter volumes in the DMPFC (e.g., owing to compensatory, neuroplastic remodelling mechanisms) needs to be confirmed in future studies.

Better tolerability of St. John's wort extract WS 5570 compared to treatment with SSRIs: a reanalysis of data from controlled clinical trials in acute major depression.

Selective serotonin reuptake inhibitors (SSRIs) like paroxetine have replaced older antidepressants mainly because of a more favorable safety profile, but they are still associated with burdensome side effects. We investigate the tolerability of St. John's wort extract WS 5570, a herbal antidepressant with proven efficacy, in comparison to paroxetine and other SSRIs and placebo. A reanalysis was performed based on the original data from four controlled clinical trials during which 1661 outpatients with major depression (Diagnostic and Statistical Manual of Mental Disorders, 4th Edition criteria) received between 600 and 1800 mg/day WS 5570 (n=1264), 20 or 40 mg/day paroxetine (n=126), or placebo (n=271) for 6 weeks. For single and grouped adverse events, the risk ratios for treatment group comparisons were determined along with their 95% confidence intervals, including comparisons with published data for SSRIs. Across the four trials, the percentage of patients with any adverse events under WS 5570 exposition was comparable with placebo [risk ratio (95% confidence interval): 1.1 (0.9-1.3) in favor of WS 5570] and significantly lower than for paroxetine [2.4 (2.1-2.8)]. Compared with the herbal extract adverse event rates under paroxetine were between 10 and 38-fold higher (point estimates) in five out of seven symptom clusters inspected. WS 5570 was devoid of effects of sedation, anticholinergic reactions, gastrointestinal disturbances, and sexual dysfunction often found during treatment with SSRIs and other synthetic antidepressants. In conclusion, WS 5570 exhibits substantially lower incidence rates of adverse events than paroxetine and other SSRIs.

Influence of age on effectiveness and tolerability of electroconvulsive therapy.

OBJECTIVES: The effectiveness of electroconvulsive therapy (ECT) in pharmacotherapy-resistant major depressive disorder and schizophrenia has been shown for all age groups. Nevertheless, age-specific adverse effects such as greater cognitive impairment and higher somatic risks due to medical comorbidities and concomitant medication may be limiting factors in geriatric patients. METHODS: We retrospectively evaluated 4457 treatments in 380 patients to investigate the influence of age on ECT outcome, safety, and adverse effects. Clinical variables, treatment modalities, and neurophysiological parameters were analyzed. For modeling the influence of age on these variables of interest, linear and logistic regression models were performed. RESULTS: The mean (SD) age of our patients was 51.2 (15) years; 30% were older than 60 years. Diagnoses were major depressive disorder in 74.4% and schizophrenia in 25.6%. We found a considerable clinical improvement in all age groups. A higher severity of disease at admission corresponded to a better clinical response. Analyzing treatment modalities of elderly patients older than 60 years, no significant differences in need and number of concomitant psychotropic medications were seen, but significant differences were seen in medical co-medication. Ictal and postictal neurophysiological parameters were only in part predictive for clinical outcome, but age had a significant influence on most of them. Transient cardiovascular adverse effects and cognitive disturbances were more frequent in the elderly. In most cases, there was no need for any specific treatment. CONCLUSIONS: Our data confirm previous studies indicating the good effectiveness of ECT irrespective of age. We also found an excellent tolerability profile in the elderly in our patient sample. There was no mortality, and only transient and no life-threatening adverse events occurred.

A genetic variant of HTR2C may play a role in the manifestation of Tourette syndrome.

Gilles de la Tourette syndrome (GTS) (MIM 137580) is a complex neuropsychiatric disorder probably originating from a disturbed interplay of several neurotransmitter systems in the prefrontal-limbic-basal ganglia loop. Polygenetic multifactorial inheritance has been postulated; nevertheless, no confirmed susceptible genes have been identified yet. As neuroimaging studies allude to dopaminergic and serotonergic dysfunction in GTS and serotonin as an important factor for dopamine release, genotyping of common polymorphisms in the serotonergic receptor (HTR1A: C-1019G; HTR2A: T102C, His452Tyr, A-1438G; HTR2C: C-759T, G-697C) and transporter genes (SLC6A4) was carried out in 87 patients with GTS, compared with 311 matched controls. We found a nominally significant association between both polymorphisms in the HTR2C and the GTS, which was more pronounced in male patients. Analysis of the further serotonergic polymorphisms did not reveal any significant result. A modified function of these promoter polymorphisms may contribute to the complex interplay of serotonin and dopamine and then to the manifestation of GTS.

Alzheimer disease: functional abnormalities in the dorsal visual pathway.

PURPOSE: To evaluate whether patients with Alzheimer disease (AD) have altered activation compared with age-matched healthy control (HC) subjects during a task that typically recruits the dorsal visual pathway. MATERIALS AND METHODS: The study was performed in accordance with the Declaration of Helsinki, with institutional ethics committee approval, and all subjects provided written informed consent. Two tasks were performed to investigate neural function: face matching and location matching. Twelve patients with mild AD and 14 age-matched HC subjects were included. Brain activation was measured by using functional magnetic resonance imaging. Group statistical analyses were based on a mixed-effects model corrected for multiple comparisons. RESULTS: Task performance was not statistically different between the two groups, and within groups there were no differences in task performance. In the HC group, the visual perception tasks selectively activated the visual pathways. Conversely in the AD group, there was no selective activation during performance of these same tasks. Along the dorsal visual pathway, the AD group recruited additional regions, primarily in the parietal and frontal lobes, for the location-matching task. There were no differences in activation between groups during the face-matching task. CONCLUSION: The increased activation in the AD group may represent a compensatory mechanism for decreased processing effectiveness in early visual areas of patients with AD. The findings support the idea that the dorsal visual pathway is more susceptible to putative AD-related neuropathologic changes than is the ventral visual pathway.

Expression of N-methyl-d-aspartate (NMDA) receptor subunits and splice variants in an animal model of long-term voluntary alcohol self-administration.

Long-term, free-choice, alcohol self-administration with repeated alcohol deprivation phases is known to enhance N-methyl-d-aspartate (NMDA) receptor activity. We hypothesized that this might not only reflect an increase in NMDA receptor density, but that differential transcriptional regulation and alternative splicing of the various subunits comprising the NMDA receptor may lead to changes in receptor composition and subsequent function. We, therefore, aimed to further investigate this effect in various brain regions. The relative mRNA expression of exon 5 inclusion/exclusion variants of the NR1 subunit, and the relative expression of NR2A, NR2B and NR2C subunits was examined in rats subjected to long-term free-choice, alcohol self-administration with repeated alcohol deprivation phases. We observed a relative decrease of the NR2C/NR2A mRNA ratio and an increase of NR1 splice variants including exon 5 (NR1+E5) in the striatum but not in the cortex, hippocampus or cerebellum in the experimental group. Our results demonstrate that long-term voluntary alcohol self-administration, affects the regulation of genes encoding the various subunits and splice variants of the NMDA receptor in a brain regional-specific manner.

Reliable manual segmentation of the frontal, parietal, temporal, and occipital lobes on magnetic resonance images of healthy subjects.

BACKGROUND: It is a challenge to reliably measure the lobar volumes from magnetic resonance imaging (MRI) data. OBJECTIVE: Description of a landmark-based method for volumetric segmentation of the brain into the four cerebral lobes from MR images. METHOD: The segmentation method relies on a combination of anatomical landmarks and geometrical definitions. The first step, described previously, is a segmentation of the four lobes on the surface of the brain. The internal borders between the lobes are defined on the axial slices of the brain. The intra- and inter- rater reliability was determined from the MRI scans of a group of 10 healthy control subjects measured by 2 independent raters. RESULTS: The intra-rater relative error (and intra-class correlation coefficient) of the lobar volume measures ranged from 0.81% to 3.85% (from 0.97 to 0.99). The inter-rater relative error (and intra-class correlation coefficient) ranged from 0.55% to 3.09% (from 0.94 to 0.99). CONCLUSION: This technique has been shown to have high intra- and inter-rater reliability. The current method provides a method to obtain volumetric estimates of the 4 cerebral lobes.