In the current study, we compared the changes in brain microstructure over 1 year in male adolescents with early-onset and persistent BD versus HC. We were especially interested in examining the relationship between abnormal brain function and structural changes of the thalamus and anterior cingulate cortex (ACC). Prior work found adolescent BD and BD adults to have significant reductions in thalamus and medial prefrontal cortical volume compared to HC ([@bb0035]), which could be related to altered thalamic connectivity ([@bb0030]). The thalamus is involved in multiple brain networks and plays a key role in the regulation of sleep, body temperature, and attentional performance ([@bb0020]), as well as coordinating sensory, cognitive, and emotional signals into goal-directed actions ([@bb0095]). There is considerable evidence for a dysfunction of the thalamus in adults with BD ([@bb0045]) and a relationship between thalamic volume and cognitive performance ([@bb0010]). The anterior cingulate cortex (ACC) also has a role in attention, as it is one of the nodes in the salience network which plays a crucial role in switching and regulating between different brain networks ([@bb0065], [@bb0070]). This network helps to determine the allocation of attention to a stimulus and is important in the process of emotion regulation ([@bb0070]). Aims {#s0010} ---- The aims of this study were to (a) use longitudinal analysis to identify differences in thalamic and ACC volumes between adolescents with BD versus HC, and (b) to examine the association between subcortical and cortical volumetric changes and clinical characteristics. Methods {#s0015} ======= Participants {#s0020} ------------ Sixty-four adolescents were included in the study (34 with BD and 30 HC). The mean age was 14.4 (SD = 1.2) for the BD group and 14.3 (SD = 1.2) for the HC group. A significant difference in age between the two groups was observed (*p* \< .05), with a trend towards significance when age was tested for the interaction of diagnosis and time (*p* = .06). Demographic data are shown in [Table 1](#t0005){ref-type="table"}. The age of onset of BD ranged from 11 to 15 years old (13.57; SD = 1.2) with a mean duration of 1.3 years (SD = 0.6). All participants were recruited from the community and assessed in clinic. Participants were excluded from the study if they were diagnosed with a neurological disorder (e.g., traumatic brain injury, cerebral tumor, stroke, multiple sclerosis, epilepsy, metabolic disorder, or intellectual disability). Participants were also excluded if they were prescribed a psychotropic medication that could affect brain structure. All participants had a full-scale Wechsler Intelligence Scale for Children (WISC-III) ([@bb0075]) or Wechsler Adult Intelligence Scale for adults (WAIS-III) ([@bb0085]) for cognitive testing and the Wechsler Test of Adult Reading (WTAR) ([@bb0080]) for age calculation. Participants were also assessed using the Wender Utah Rating Scale (WURS) ([@bb0075]), the Young Mania Rating Scale (YMRS) ([@bb0090]), the Hamilton Depression Rating Scale (HAM-D) ([@bb0070]), and the Child Behavior Checklist (CBCL) ([@bb0035]). Participants did not undergo a structured clinical interview for major depressive disorder (MDD) because they are at an increased risk for suicide ([@bb0075]). Parents completed the CBCL as well. A control group of age and gender-matched HC was recruited from the community for comparison. The HCs were included if they had no history of Axis I disorders (except for substance abuse or dependence), as assessed by the Schedule for Affective Disorders and Schizophrenia for School-Age Children-Present and Lifetime Version (K-SADS-PL), and did not report any history of a medical disorder that could impact brain development or affect neurocognitive function (e.g., cerebral palsy, stroke, diabetes, arthritis, or attention deficit hyperactivity disorder). There were no statistically significant differences in socioeconomic status or ethnicity between the BD and HC groups. The study was approved by the human research ethics committee of the University of New South Wales and the Australian Catholic University. Written informed consent was obtained from the adolescents and their parents. T1 Structural Magnetic Resonance Imaging (MRI) Protocol {#s0025} ------------------------------------------------------- MRI data were acquired using a 1.5 Tesla GE Signa whole body scanner at the Brain and Mind Research Institute. Three-dimensional fast spoiled gradient-recalled (3D-FSPGR) T1 structural scans were acquired using the following protocol: sagittal 3D-T1 SNAP with inversion recovery (IR), TR = 22 ms, TE = 5.9 ms, TI = 450 ms, field of view 256 mm, matrix = 256 × 256, slices = 160, slice thickness 1.5 mm, NEX = 1, flip angle = 15°. MRI Analysis {#s0030} ------------ Images were analysed with SPM8 (Wellcome Department of Cognitive Neurology, London, UK). The images were manually co-registered and segmented into grey matter, white matter, and cerebrospinal fluid (CSF) using the T1 segmented images. Cortical and subcortical structures were segmented using the unified segmentation algorithm ([@bb0125]) and then spatially normalised to the MNI space. The final spatial normalised images were smoothed with a Gaussian kernel of 8 mm full-width at half maximum (FWHM) ([@bb0075]). Statistics {#s0035} ---------- Statistical analysis was performed using the SPSS software program (version 21, IBM SPSS Statistics, IBM Corporation, Armonk, NY). To assess differences in volumetric changes over 1 year between the BD and HC groups, a repeated measures ANOVA model (Mixed-effects) was used to determine the main effects of time and diagnosis, as well as interactions of these variables with covariates of age, medication use, gender, and duration of illness. The Bonferroni post-hoc tests were used to examine significant interactions. For significant findings, to identify the direction of effects we investigated the nature of the interaction using one-way ANOVA to compare groups on the mean of the values at different time points. All significance was set to *p* \< .05. Correlation analysis was performed using Spearman\'s *r* to examine the association between clinical outcomes and changes in volume over 1 year. A bivariate correlation test was also performed to explore the relationship between clinical variables and changes in volume over 1 year. Age was not used as a covariate in either analysis because it was the dependent variable. Results {#s0040} ======= [Table 2](#t0010){ref-type="table"} shows the mean and standard deviations of volume change over 1 year in the thalamus and ACC of participants with BD and HC. The volume of the thalamus was significantly smaller in the BD group compared to HC over 1 year (F = 5.9, *p* \< .05). A repeated measures ANOVA model showed that there were no significant differences in thalamic volume over 1 year in the HC group. In the BD group, thalamic volume decreased over time (F = 6.6, *p* \< .01). When correlations between the clinical characteristics and thalamic volume over 1 year in the BD group were examined, thalamic volume was negatively correlated with the HAM-D (*r* = − 0.53, *p* \< .05) and the HAM-A (*r* = − 0.41, *p* \< .05), suggesting that as symptoms of depression and anxiety increased, thalamic volume decreased. No significant correlations between other clinical variables and changes in thalamic volume were found. No significant changes in ACC volume over 1 year were found in either the BD or HC group. A repeated measures ANOVA showed no differences in ACC volume over time (F = 0.5, *p* \> .05). When analysis was repeated without the BD adolescents who were using medication at the baseline MRI scan, results showed that thalamic volume was significantly smaller in the BD group compared to HC over 1 year (F = 4.8, *p* \< .01). The BD group showed a significant decrease in thalamic volume over 1 year (F = 7.9, *p* \< .05). For the BD group, the interaction of diagnosis and time was no longer significant. However, the main effect of diagnosis (F = 10.4, *p* \< .001) and the main effect of time (F = 12.7, *p* \< .001) remained significant. Discussion {#s0045} ========== This study found smaller thalamic volume in BD compared to HC over 1 year. The BD group showed a significant decrease in thalamic volume over 1 year. Thalamic volume was also negatively correlated with symptom severity as indexed by the HAM-D and HAM-A scales, suggesting that smaller thalamic volume was associated with greater symptoms of depression and anxiety. Thalamic volume also significantly decreased over 1 year in the BD group, independent of any medication effects on brain structure