Accurate diagnosis of bipolar disorder (BPD) is difficult in clinical practice, with an average delay between symptom onset and diagnosis of about 7 years. A depressive episode often precedes the first manic episode, making it difficult to distinguish BPD from unipolar major depressive disorder (MDD).

We use genome-wide association analyses (GWAS) to identify differential genetic factors and to develop predictors based on polygenic risk scores (PRS) that may aid early differential diagnosis.

Based on individual genotypes from case–control cohorts of BPD and MDD shared through the Psychiatric Genomics Consortium, we compile case–case–control cohorts, applying a careful quality control procedure. In a resulting cohort of 51 149 individuals (15 532 BPD patients, 12 920 MDD patients and 22 697 controls), we perform a variety of GWAS and PRS analyses.

Although our GWAS is not well powered to identify genome-wide significant loci, we find significant chip heritability and demonstrate the ability of the resulting PRS to distinguish BPD from MDD, including BPD cases with depressive onset (BPD-D). We replicate our PRS findings in an independent Danish cohort (iPSYCH 2015, N = 25 966). We observe strong genetic correlation between our case–case GWAS and that of case–control BPD.

We find that MDD and BPD, including BPD-D are genetically distinct. Our findings support that controls, MDD and BPD patients primarily lie on a continuum of genetic risk. Future studies with larger and richer samples will likely yield a better understanding of these findings and enable the development of better genetic predictors distinguishing BPD and, importantly, BPD-D from MDD.

Maternal vitamin-D and omega-3 fatty acid (DHA) deficiencies during pregnancy have previously been associated with offspring neurodevelopmental traits. However, observational study designs cannot distinguish causal effects from confounding.

First, we conducted Mendelian randomisation (MR) using genetic instruments for vitamin-D and DHA identified in independent genome-wide association studies (GWAS). Outcomes were (1) GWAS for traits related to autism and ADHD, generated in the Norwegian mother, father, and child cohort study (MoBa) from 3 to 8 years, (2) autism and ADHD diagnoses. Second, we used mother–father–child trio-MR in MoBa (1) to test causal effects through maternal nutrient levels, (2) to test effects of child nutrient levels, and (3) as a paternal negative control.

Associations between higher maternal vitamin-D levels on lower ADHD related traits at age 5 did not remain after controlling for familial genetic predisposition using trio-MR. Furthermore, we did not find evidence for causal maternal effects of vitamin-D/DHA levels on other offspring traits or diagnoses. In the reverse direction, there was evidence for a causal effect of autism genetic predisposition on lower vitamin-D levels and of ADHD genetic predisposition on lower DHA levels.

Triangulating across study designs, we did not find evidence for maternal effects. We add to a growing body of evidence that suggests that previous observational associations are likely biased by genetic confounding. Consequently, maternal supplementation is unlikely to influence these offspring neurodevelopmental traits. Notably, genetic predisposition to ADHD and autism was associated with lower DHA and vitamin-D levels respectively, suggesting previous associations might have been due to reverse causation.

Cognitive training is a non-pharmacological intervention aimed at improving cognitive function across a single or multiple domains. Although the underlying mechanisms of cognitive training and transfer effects are not well-characterized, cognitive training has been thought to facilitate neural plasticity to enhance cognitive performance. Indeed, the Scaffolding Theory of Aging and Cognition (STAC) proposes that cognitive training may enhance the ability to engage in compensatory scaffolding to meet task demands and maintain cognitive performance. We therefore evaluated the effects of cognitive training on working memory performance in older adults without dementia. This study will help begin to elucidate non-pharmacological intervention effects on compensatory scaffolding in older adults.

48 participants were recruited for a Phase III randomized clinical trial (Augmenting Cognitive Training in Older Adults [ACT]; NIH R01AG054077) conducted at the University of Florida and University of Arizona. Participants across sites were randomly assigned to complete cognitive training (n=25) or an education training control condition (n=23). Cognitive training and the education training control condition were each completed during 60 sessions over 12 weeks for 40 hours total. The education training control condition involved viewing educational videos produced by the National Geographic Channel. Cognitive training was completed using the Posit Science Brain HQ training program, which included 8 cognitive training paradigms targeting attention/processing speed and working memory. All participants also completed demographic questionnaires, cognitive testing, and an fMRI 2-back task at baseline and at 12-weeks following cognitive training.

Repeated measures analysis of covariance (ANCOVA), adjusted for training adherence, transcranial direct current stimulation (tDCS) condition, age, sex, years of education, and Wechsler Test of Adult Reading (WTAR) raw score, revealed a significant 2-back by training group interaction (F[1,40]=6.201, p=.017, η2=.134). Examination of simple main effects revealed baseline differences in 2-back performance (F[1,40]=.568, p=.455, η2=.014). After controlling for baseline performance, training group differences in 2-back performance was no longer statistically significant (F[1,40]=1.382, p=.247, η2=.034).

After adjusting for baseline performance differences, there were no significant training group differences in 2-back performance, suggesting that the randomization was not sufficient to ensure adequate distribution of participants across groups. Results may indicate that cognitive training alone is not sufficient for significant improvement in working memory performance on a near transfer task. Additional improvement may occur with the next phase of this clinical trial, such that tDCS augments the effects of cognitive training and results in enhanced compensatory scaffolding even within this high performing cohort. Limitations of the study include a highly educated sample with higher literacy levels and the small sample size was not powered for transfer effects analysis. Future analyses will include evaluation of the combined intervention effects of a cognitive training and tDCS on nback performance in a larger sample of older adults without dementia.

Cognitive training using a visual speed-of-processing task, called the Useful Field of View (UFOV) task, reduced dementia risk and reduced decline in activities of daily living at a 10-year follow-up in older adults. However, there is variability in the level of cognitive gains after cognitive training across studies. One potential explanation for this variability could be moderating factors. Prior studies suggest variables moderating cognitive training gains share features of the training task. Learning trials of the Hopkins Verbal Learning Test-Revised (HVLT-R) and Brief Visuospatial Memory Test-Revised (BVMT-R) recruit similar cognitive abilities and have overlapping neural correlates with the UFOV task and speed-ofprocessing/working memory tasks and therefore could serve as potential moderators. Exploring moderating factors of cognitive training gains may boost the efficacy of interventions, improve rigor in the cognitive training literature, and eventually help provide tailored treatment recommendations. This study explored the association between the HVLT-R and BVMT-R learning and the UFOV task, and assessed the moderation of HVLT-R and BVMT-R learning on UFOV improvement after a 3-month speed-ofprocessing/attention and working memory cognitive training intervention in cognitively healthy older adults.

75 healthy older adults (M age = 71.11, SD = 4.61) were recruited as part of a larger clinical trial through the Universities of Florida and Arizona. Participants were randomized into a cognitive training (n=36) or education control (n=39) group and underwent a 40-hour, 12-week intervention. Cognitive training intervention consisted of practicing 4 attention/speed-of-processing (including the UFOV task) and 4 working memory tasks. Education control intervention consisted of watching 40-minute educational videos. The HVLT-R and BVMT-R were administered at the pre-intervention timepoint as part of a larger neurocognitive battery. The learning ratio was calculated as: trial 3 total - trial 1 total/12 - trial 1 total. UFOV performance was measured at pre- and post-intervention time points via the POSIT Brain HQ Double Decision Assessment. Multiple linear regressions predicted baseline Double Decision performance from HVLT-R and BVMT-R learning ratios controlling for study site, age, sex, and education. A repeated measures moderation analysis assessed the moderation of HVLT-R and BVMT-R learning ratio on Double Decision change from pre- to post-intervention for cognitive training and education control groups.

Baseline Double Decision performance significantly associated with BVMT-R learning ratio (β=-.303, p=.008), but not HVLT-R learning ratio (β=-.142, p=.238). BVMT-R learning ratio moderated gains in Double Decision performance (p<.01); for each unit increase in BVMT-R learning ratio, there was a .6173 unit decrease in training gains. The HVLT-R learning ratio did not moderate gains in Double Decision performance (p>.05). There were no significant moderations in the education control group.

Better visuospatial learning was associated with faster Double Decision performance at baseline. Those with poorer visuospatial learning improved most on the Double Decision task after training, suggesting that healthy older adults who perform below expectations may show the greatest training gains. Future cognitive training research studying visual speed-of-processing interventions should account for differing levels of visuospatial learning at baseline, as this could impact the magnitude of training outcomes.

Aging is associated with disruptions in functional connectivity within the default mode (DMN), frontoparietal control (FPCN), and cingulo-opercular (CON) resting-state networks. Greater within-network connectivity predicts better cognitive performance in older adults. Therefore, strengthening network connectivity, through targeted intervention strategies, may help prevent age-related cognitive decline or progression to dementia. Small studies have demonstrated synergistic effects of combining transcranial direct current stimulation (tDCS) and cognitive training (CT) on strengthening network connectivity; however, this association has yet to be rigorously tested on a large scale. The current study leverages longitudinal data from the first-ever Phase III clinical trial for tDCS to examine the efficacy of an adjunctive tDCS and CT intervention on modulating network connectivity in older adults.

This sample included 209 older adults (mean age = 71.6) from the Augmenting Cognitive Training in Older Adults multisite trial. Participants completed 40 hours of CT over 12 weeks, which included 8 attention, processing speed, and working memory tasks. Participants were randomized into active or sham stimulation groups, and tDCS was administered during CT daily for two weeks then weekly for 10 weeks. For both stimulation groups, two electrodes in saline-soaked 5x7 cm2 sponges were placed at F3 (cathode) and F4 (anode) using the 10-20 measurement system. The active group received 2mA of current for 20 minutes. The sham group received 2mA for 30 seconds, then no current for the remaining 20 minutes.

Participants underwent resting-state fMRI at baseline and post-intervention. CONN toolbox was used to preprocess imaging data and conduct region of interest (ROI-ROI) connectivity analyses. The Artifact Detection Toolbox, using intermediate settings, identified outlier volumes. Two participants were excluded for having greater than 50% of volumes flagged as outliers. ROI-ROI analyses modeled the interaction between tDCS group (active versus sham) and occasion (baseline connectivity versus postintervention connectivity) for the DMN, FPCN, and CON controlling for age, sex, education, site, and adherence.

Compared to sham, the active group demonstrated ROI-ROI increases in functional connectivity within the DMN following intervention (left temporal to right temporal [T(202) = 2.78, pFDR < 0.05] and left temporal to right dorsal medial prefrontal cortex [T(202) = 2.74, pFDR < 0.05]. In contrast, compared to sham, the active group demonstrated ROI-ROI decreases in functional connectivity within the FPCN following intervention (left dorsal prefrontal cortex to left temporal [T(202) = -2.96, pFDR < 0.05] and left dorsal prefrontal cortex to left lateral prefrontal cortex [T(202) = -2.77, pFDR < 0.05]). There were no significant interactions detected for CON regions.

These findings (a) demonstrate the feasibility of modulating network connectivity using tDCS and CT and (b) provide important information regarding the pattern of connectivity changes occurring at these intervention parameters in older adults. Importantly, the active stimulation group showed increases in connectivity within the DMN (a network particularly vulnerable to aging and implicated in Alzheimer’s disease) but decreases in connectivity between left frontal and temporal FPCN regions. Future analyses from this trial will evaluate the association between these changes in connectivity and cognitive performance post-intervention and at a one-year timepoint.

Nonpathological aging has been linked to decline in both verbal and visuospatial memory abilities in older adults. Disruptions in resting-state functional connectivity within well-characterized, higherorder cognitive brain networks have also been coupled with poorer memory functioning in healthy older adults and in older adults with dementia. However, there is a paucity of research on the association between higherorder functional connectivity and verbal and visuospatial memory performance in the older adult population. The current study examines the association between resting-state functional connectivity within the cingulo-opercular network (CON), frontoparietal control network (FPCN), and default mode network (DMN) and verbal and visuospatial learning and memory in a large sample of healthy older adults. We hypothesized that greater within-network CON and FPCN functional connectivity would be associated with better immediate verbal and visuospatial memory recall. Additionally, we predicted that within-network DMN functional connectivity would be associated with improvements in delayed verbal and visuospatial memory recall. This study helps to glean insight into whether within-network CON, FPCN, or DMN functional connectivity is associated with verbal and visuospatial memory abilities in later life.

330 healthy older adults between 65 and 89 years old (mean age = 71.6 ± 5.2) were recruited at the University of Florida (n = 222) and the University of Arizona (n = 108). Participants underwent resting-state fMRI and completed verbal memory (Hopkins Verbal Learning Test - Revised [HVLT-R]) and visuospatial memory (Brief Visuospatial Memory Test - Revised [BVMT-R]) measures. Immediate (total) and delayed recall scores on the HVLT-R and BVMT-R were calculated using each test manual’s scoring criteria. Learning ratios on the HVLT-R and BVMT-R were quantified by dividing the number of stimuli (verbal or visuospatial) learned between the first and third trials by the number of stimuli not recalled after the first learning trial. CONN Toolbox was used to extract average within-network connectivity values for CON, FPCN, and DMN. Hierarchical regressions were conducted, controlling for sex, race, ethnicity, years of education, number of invalid scans, and scanner site.

Greater CON connectivity was significantly associated with better HVLT-R immediate (total) recall (ß = 0.16, p = 0.01), HVLT-R learning ratio (ß = 0.16, p = 0.01), BVMT-R immediate (total) recall (ß = 0.14, p = 0.02), and BVMT-R delayed recall performance (ß = 0.15, p = 0.01). Greater FPCN connectivity was associated with better BVMT-R learning ratio (ß = 0.13, p = 0.04). HVLT-R delayed recall performance was not associated with connectivity in any network, and DMN connectivity was not significantly related to any measure.

Connectivity within CON demonstrated a robust relationship with different components of memory function as well across verbal and visuospatial domains. In contrast, FPCN only evidenced a relationship with visuospatial learning, and DMN was not significantly associated with memory measures. These data suggest that CON may be a valuable target in longitudinal studies of age-related memory changes, but also a possible target in future non-invasive interventions to attenuate memory decline in older adults.

Studying phenotypic and genetic characteristics of age at onset (AAO) and polarity at onset (PAO) in bipolar disorder can provide new insights into disease pathology and facilitate the development of screening tools.

To examine the genetic architecture of AAO and PAO and their association with bipolar disorder disease characteristics.

Genome-wide association studies (GWASs) and polygenic score (PGS) analyses of AAO (n = 12 977) and PAO (n = 6773) were conducted in patients with bipolar disorder from 34 cohorts and a replication sample (n = 2237). The association of onset with disease characteristics was investigated in two of these cohorts.

Earlier AAO was associated with a higher probability of psychotic symptoms, suicidality, lower educational attainment, not living together and fewer episodes. Depressive onset correlated with suicidality and manic onset correlated with delusions and manic episodes. Systematic differences in AAO between cohorts and continents of origin were observed. This was also reflected in single-nucleotide variant-based heritability estimates, with higher heritabilities for stricter onset definitions. Increased PGS for autism spectrum disorder (β = −0.34 years, s.e. = 0.08), major depression (β = −0.34 years, s.e. = 0.08), schizophrenia (β = −0.39 years, s.e. = 0.08), and educational attainment (β = −0.31 years, s.e. = 0.08) were associated with an earlier AAO. The AAO GWAS identified one significant locus, but this finding did not replicate. Neither GWAS nor PGS analyses yielded significant associations with PAO.

AAO and PAO are associated with indicators of bipolar disorder severity. Individuals with an earlier onset show an increased polygenic liability for a broad spectrum of psychiatric traits. Systematic differences in AAO across cohorts, continents and phenotype definitions introduce significant heterogeneity, affecting analyses.

Observational studies have found associations between smoking and both poorer cognitive ability and lower educational attainment; however, evaluating causality is challenging. We used two complementary methods to explore this.

We conducted observational analyses of up to 12 004 participants in a cohort study (Study One) and Mendelian randomisation (MR) analyses using summary and cohort data (Study Two). Outcome measures were cognitive ability at age 15 and educational attainment at age 16 (Study One), and educational attainment and fluid intelligence (Study Two).

Study One: heaviness of smoking at age 15 was associated with lower cognitive ability at age 15 and lower educational attainment at age 16. Adjustment for potential confounders partially attenuated findings (e.g. fully adjusted cognitive ability β −0.736, 95% CI −1.238 to −0.233, p = 0.004; fully adjusted educational attainment β −1.254, 95% CI −1.597 to −0.911, p < 0.001). Study Two: MR indicated that both smoking initiation and lifetime smoking predict lower educational attainment (e.g. smoking initiation to educational attainment inverse-variance weighted MR β −0.197, 95% CI −0.223 to −0.171, p = 1.78 × 10−49). Educational attainment results were robust to sensitivity analyses, while analyses of general cognitive ability were less so.

We find some evidence of a causal effect of smoking on lower educational attainment, but not cognitive ability. Triangulation of evidence across observational and MR methods is a strength, but the genetic variants associated with smoking initiation may be pleiotropic, suggesting caution in interpreting these results. The nature of this pleiotropy warrants further study.

Schizophrenia is a highly heritable disorder with undetermined neurobiological causes. Understanding the impact on brain anatomy of carrying genetic risk for the disorder will contribute to uncovering its neurobiological underpinnings.

To examine the effect of rare copy number variants (CNVs) associated with schizophrenia on brain cortical anatomy in a sample of unaffected participants from the UK Biobank.

We used regression analyses to compare cortical thickness and surface area (total and across gyri) between 120 unaffected carriers of rare CNVs associated with schizophrenia and 16 670 participants without any pathogenic CNV. A measure of cortical thickness and surface area covariance across gyri was also compared between groups.

Carrier status was associated with reduced surface area (β = −0.020 mm2, P < 0.001) and less robustly with increased cortical thickness (β = 0.015 mm, P = 0.035), and with increased covariance in thickness (carriers z = 0.31 v. non-carriers z = 0.22, P < 0.0005). Associations were mainly present in frontal and parietal areas and driven by a limited number of rare risk alleles included in our analyses (mainly 15q11.2 deletion for surface area and 16p13.11 duplication for thickness covariance).

Results for surface area conformed with previous clinical findings, supporting surface area reductions as an indicator of genetic liability for schizophrenia. Results for cortical thickness, though, argued against its validity as a potential risk marker. Increased structural thickness covariance across gyri also appears related to risk for schizophrenia. The heterogeneity found across the effects of rare risk alleles suggests potential different neurobiological gateways into schizophrenia's phenotype.

It is not clear to what extent associations between schizophrenia, cannabis use and cigarette use are due to a shared genetic etiology. We, therefore, examined whether schizophrenia genetic risk associates with longitudinal patterns of cigarette and cannabis use in adolescence and mediating pathways for any association to inform potential reduction strategies.

Associations between schizophrenia polygenic scores and longitudinal latent classes of cigarette and cannabis use from ages 14 to 19 years were investigated in up to 3925 individuals in the Avon Longitudinal Study of Parents and Children. Mediation models were estimated to assess the potential mediating effects of a range of cognitive, emotional, and behavioral phenotypes.

The schizophrenia polygenic score, based on single nucleotide polymorphisms meeting a training-set p threshold of 0.05, was associated with late-onset cannabis use (OR = 1.23; 95% CI = 1.08,1.41), but not with cigarette or early-onset cannabis use classes. This association was not mediated through lower IQ, victimization, emotional difficulties, antisocial behavior, impulsivity, or poorer social relationships during childhood. Sensitivity analyses adjusting for genetic liability to cannabis or cigarette use, using polygenic scores excluding the CHRNA5-A3-B4 gene cluster, or basing scores on a 0.5 training-set p threshold, provided results consistent with our main analyses.

Our study provides evidence that genetic risk for schizophrenia is associated with patterns of cannabis use during adolescence. Investigation of pathways other than the cognitive, emotional, and behavioral phenotypes examined here is required to identify modifiable targets to reduce the public health burden of cannabis use in the population.

Previous literature has demonstrated a strong association between cigarette smoking, suicidal ideation and suicide attempts. This association has not previously been examined in a causal inference framework and could have important implications for suicide prevention strategies.

We aimed to examine the evidence for an association between smoking behaviours (initiation, smoking status, heaviness, lifetime smoking) and suicidal thoughts or attempts by triangulating across observational and Mendelian randomisation analyses.

First, in the UK Biobank, we calculated observed associations between smoking behaviours and suicidal thoughts or attempts. Second, we used Mendelian randomisation to explore the relationship between smoking and suicide attempts and ideation, using genetic variants as instruments to reduce bias from residual confounding and reverse causation.

Our observational analysis showed a relationship between smoking behaviour, suicidal ideation and attempts, particularly between smoking initiation and suicide attempts (odds ratio, 2.07; 95% CI 1.91–2.26; P < 0.001). The Mendelian randomisation analysis and single-nucleotide polymorphism analysis, however, did not support this (odds ratio for lifetime smoking on suicidal ideation, 0.050; 95% CI −0.027 to 0.127; odds ratio on suicide attempts, 0.053; 95% CI, −0.003 to 0.110). Despite past literature showing a positive dose-response relationship, our results showed no clear evidence for a causal effect of smoking on suicidal ideation or attempts.

This was the first Mendelian randomisation study to explore the effect of smoking on suicidal ideation and attempts. Our results suggest that, despite observed associations, there is no clear evidence for a causal effect.

Smoking prevalence is higher amongst individuals with schizophrenia and depression compared with the general population. Mendelian randomisation (MR) can examine whether this association is causal using genetic variants identified in genome-wide association studies (GWAS).

We conducted two-sample MR to explore the bi-directional effects of smoking on schizophrenia and depression. For smoking behaviour, we used (1) smoking initiation GWAS from the GSCAN consortium and (2) we conducted our own GWAS of lifetime smoking behaviour (which captures smoking duration, heaviness and cessation) in a sample of 462690 individuals from the UK Biobank. We validated this instrument using positive control outcomes (e.g. lung cancer). For schizophrenia and depression we used GWAS from the PGC consortium.

There was strong evidence to suggest smoking is a risk factor for both schizophrenia (odds ratio (OR) 2.27, 95% confidence interval (CI) 1.67–3.08, p < 0.001) and depression (OR 1.99, 95% CI 1.71–2.32, p < 0.001). Results were consistent across both lifetime smoking and smoking initiation. We found some evidence that genetic liability to depression increases smoking (β = 0.091, 95% CI 0.027–0.155, p = 0.005) but evidence was mixed for schizophrenia (β = 0.022, 95% CI 0.005–0.038, p = 0.009) with very weak evidence for an effect on smoking initiation.

These findings suggest that the association between smoking, schizophrenia and depression is due, at least in part, to a causal effect of smoking, providing further evidence for the detrimental consequences of smoking on mental health.