At baseline, the researchers found that the coherence of neural activity within the brain’s default mode network was greater in persons with dysthymia than in healthy controls. Following the 10-week trial, they found that treatment with duloxetine, but not placebo, normalized default mode network connectivity (P < .03). “If they received placebo, the activity [in this brain region] didn’t change at all; it’s exactly the same as it was before the medication trial,” Dr. Peterson said. “If they received active medication, activity normalized; it reduced the cross-talk across nodes of the default mode network so that now, their [default mode network] activity is no longer discernible or different from the healthy controls. This shows that duloxetine is causing the reduction in the cross-talk between these circuits, and by doing so, it’s normalizing activity in the default mode system.”
A similar effect was observed in a study that assessed the impact of stimulant medications in children with attention-deficit/hyperactivity disorder (ADHD). Specifically, researchers including Dr. Peterson used cross-sectional MRI to examine the morphologic features of the basal ganglia nuclei in 48 children with ADHD who were off medication and 56 healthy controls (Am. J. Psychiatry 2010;167:977-86).
“Reduced volume in portions of the basal ganglia structures is important for impulse control and attention,” Dr. Peterson said. “We found that those same structures are enlarged when kids are on their medication so as not to be different from healthy controls. We think that stimulant medications in ADHD are normalizing these disturbances in the structure of the basal ganglia.”
Identifying neurometabolic dysfunction
Prior studies measured lactate in peripheral blood, muscle, or postmortem samples of people with autism spectrum disorders (ASD), “but these do not necessarily indicate the presence of metabolic dysfunction in the brain,” Dr. Peterson said. In a recent study he and other researchers used magnetic resonance spectroscopic imaging to measure lactate in the brains of people at risk for ASD. The analysis included 75 high-functioning ASD participants and 96 typically developing children and adults (JAMA Psychiatry 2014;71:665:71). Definite lactate peaks were present at a significantly higher rate in the ASD participants, compared with controls (13% vs. 1%, P = .001). In addition, the presence of lactate was significantly greater in adults, compared with children (20% vs. 6%; P = .004), ”perhaps suggesting that this could be a degenerative process that exacerbates through time,” Dr. Peterson said.
The presence of lactate did not correlate with clinical symptoms based on ASD subtype, autism diagnostic observation schedule domain score, or full-scale IQ. Dr. Peterson said the presence of lactate is “definitive proof that mitochondria are dysfunctional in the brains of a substantial number of autistic people. This disturbance is found now in autism, but it will likely be true in people with other neuropsychiatric disorders as well.”
A key advantage of MR spectroscopic imaging, he continued, “is that we can determine where in the brain lactate’s being produced. These kinds of studies will help guide studies in mitochondrial genetics and dysfunction in ASD and other conditions. It also has important clinical implications, because there are novel treatment approaches now for mitochondrial dysfunction, such as dietary interventions that can reduce the metabolic dysfunction.”
Using automated, brain-based diagnostic classifications
Dr. Peterson and other researchers used an automated method to diagnose individuals as having one of various neuropsychiatric illnesses using only anatomical MRI scans. “The method employs a semisupervised learning algorithm that discovers natural groupings of brains based on the spatial patterns of variation in the morphology of the cerebral cortex and other brain regions,” they explained in their article (PLoS One 2012;7:e50698). “We used split-half and leave-one-out cross-validation analyses in large MRI datasets to assess the reproducibility and diagnostic accuracy of those groupings.”
Conceptually, “we look at the volume increases and decreases throughout the surface of the cortex,” Dr. Peterson said at the meeting. “We do the same thing at the basal ganglia, thalamus, cerebellum, amygdala, and hippocampus. What we’re trying to do is to use structural variation in the brain to identify spatial patterns in brain structure that help to identify people who have brain features in common, just as the pattern of dermatomal ridges on your finger can identify specific individuals with great accuracy.
“We use machine learning algorithms to identify a pattern of abnormality in brain structure that’s more similar among people with Tourette’s syndrome, for example, than in people who have ADHD.”
Using this automated approach to making clinical diagnoses yielded impressive results. For example, the sensitivity and specificity to differentiate ADHD subjects from healthy controls was 93.6% and 89.5%, respectively. It also was strong for diagnosis of Tourette’s syndrome, (94.6% sensitive and 79% specific) and for schizophrenia (93.1% sensitive and 94.5% specific).