A Robust and Accurate Deep-learning-based Method for the Segmentation of Subcortical Brain: Cross-dataset Evaluation of Generalization Performance.

PURPOSE: To analyze subcortical brain volume more reliably, we propose a deep learning segmentation method of subcortical brain based on magnetic resonance imaging (MRI) having high generalization performance, accuracy, and robustness. METHODS: First, local images of three-dimensional (3D) bounding boxes were extracted for seven subcortical structures (thalamus, putamen, caudate, pallidum, hippocampus, amygdala, and accumbens) from a whole brain MR image as inputs to the neural network. Second, dilated convolution layers, which input information of variable scope, were introduced to the blocks that make up the neural network. These blocks were connected in parallel to simultaneously process global and local information obtained by the dilated convolution layers. To evaluate generalization performance, different datasets were used for training and testing sessions (cross-dataset evaluation) because subcortical brain segmentation in clinical analysis is assumed to be applied to unknown datasets. RESULTS: The proposed method showed better generalization performance that can obtain stable accuracy for all structures, whereas the state-of-the-art deep learning method obtained extremely low accuracy for some structures. The proposed method performed segmentation for all samples without failing with significantly higher accuracy (P < 0.005) than conventional methods such as 3D U-Net, FreeSurfer, and Functional Magnetic Resonance Imaging of the Brain's (FMRIB's) Integrated Registration and Segmentation Tool in the FMRIB Software Library (FSL-FIRST). Moreover, when applying this proposed method to larger datasets, segmentation was robustly performed for all samples without producing segmentation results on the areas that were apparently different from anatomically relevant areas. On the other hand, FSL-FIRST produced segmentation results on the area that were apparently and largely different from the anatomically relevant area for about one-third to one-fourth of the datasets. CONCLUSION: The cross-dataset evaluation showed that the proposed method is superior to existing methods in terms of generalization performance, accuracy, and robustness.

Age-related white matter changes revealed by a whole-brain fiber-tracking method in bipolar disorder compared to major depressive disorder and healthy controls.

AIM: Several studies have reported altered age-associated changes in white matter integrity in bipolar disorder (BD). However, little is known as to whether these age-related changes are illness-specific. We assessed disease-specific effects by controlling for age and investigated age-associated changes and Group × Age interactions in white matter integrity among major depressive disorder (MDD) patients, BD patients, and healthy controls. METHODS: Healthy controls (n = 96; age range, 20-77 years), MDD patients (n = 101; age range, 25-78 years), and BD patients (n = 58; age range, 22-76 years) participated in this study. Fractional anisotropy (FA) derived from diffusion tensor imaging in 54 white matter tracts were compared after controlling for the linear and quadratic effect of age using a generalized linear model. Age-related effects and Age × Group interactions were also assessed in the model. RESULTS: The main effect of group was significant in the left column and body of the fornix after controlling for both linear and quadratic effects of age, and in the left body of the corpus callosum after controlling for the quadratic effect of age. BD patients exhibited significantly lower FA relative to other groups. There was no Age × Group interaction in the tracts. CONCLUSION: Significant FA reductions were found in BD patients after controlling for age, indicating that abnormal white matter integrity in BD may occur at a younger age rather than developing progressively with age.

White matter abnormalities and cognitive function in euthymic patients with bipolar disorder and major depressive disorder.

OBJECTIVES: In recent years, a growing number of diffusion tensor imaging (DTI) studies have compared white matter integrity between patients with major depressive disorder (MDD) and bipolar disorder (BD). However, few studies have examined the pathophysiological significance of different degrees of white matter abnormalities between the two disorders. The present study comprehensively assessed white matter integrity among healthy controls (HC) and euthymic patients with MDD and BD using whole-brain tractography and examined associations between white matter integrity and cognitive functioning. METHODS: We performed neurocognitive examinations and DTI with 30 HCs, 30 patients with MDD, and 30 patients with BD. We statistically evaluated white matter integrity and cognitive function differences across the three groups, assessing associations between white matter integrities and cognitive function. RESULTS: The BD group showed lower fractional anisotropy (FA) for the corpus callosum body, as well as lower, sustained attention and set-shifting scores compared to the other groups. FA for the left body of the corpus callosum was correlated with sustained attention in patients with BD. CONCLUSIONS: The significant reduction of white matter integrity in the corpus callosum in BD, compared to MDD, was associated with an impairment of sustained attention. This result promotes the understanding of the significance of white matter integrity in mood disorders.

Genetic influences on prefrontal activation during a verbal fluency task in children: A twin study using near-infrared spectroscopy.

OBJECTIVE: The genetic and environmental influences on prefrontal function in childhood are underinvestigated due to the difficulty of measuring prefrontal function in young subjects, for which near-infrared spectroscopy (NIRS) is a suitable functional neuroimaging technique that facilitates the easy and noninvasive measurement of blood oxygenation in the superficial cerebral cortices. METHOD: Using a two-channel NIRS arrangement, we measured changes in bilateral prefrontal blood oxygenation during a category version of the verbal fluency task (VFT) in 27 monozygotic twin pairs and 12 same-sex dizygotic twin pairs ages 5-17 years. We also assessed the participant's full-scale intelligence quotient (FIQ) and retrieved parental socioeconomic status (SES). Classical structured equation modeling was used to estimate the heritability. RESULTS: The heritability of VFT-related brain activation was estimated to be 44% and 37% in the right and left prefrontal regions, respectively. We also identified a significant genetic contribution (74%) to FIQ, but did not to VFT task performance. Parental SES was not correlated with FIQ, task performance, or task-related prefrontal activation. CONCLUSIONS: This finding provides further evidence that variance in prefrontal function has a genetic component since childhood and highlights brain function, as measured by NIRS, as a promising candidate for endophenotyping neurodevelopmental disorders.

Neuroimaging-Aided Prediction of the Effect of Methylphenidate in Children with Attention-Deficit Hyperactivity Disorder: A Randomized Controlled Trial.

Although methylphenidate hydrochloride (MPH) is a first-line treatment for children with attention-deficit hyperactivity disorder (ADHD), the non-response rate is 30%. Our aim was to develop a supplementary neuroimaging biomarker for predicting the clinical effect of continuous MPH administration by using near-infrared spectroscopy (NIRS). After baseline assessment, we performed a double-blind, placebo-controlled, crossover trial with a single dose of MPH, followed by a prospective 4-to-8-week open trial with continuous MPH administration, and an ancillary 1-year follow-up. Twenty-two drug-naïve and eight previously treated children with ADHD (NAÏVE and NON-NAÏVE) were compared with 20 healthy controls (HCs) who underwent multiple NIRS measurements without intervention. We tested whether NIRS signals at the baseline assessment or ΔNIRS (single dose of MPH minus baseline assessment) predict the Clinical Global Impressions-Severity (CGI-S) score after 4-to-8-week or 1-year MPH administration. The secondary outcomes were the effect of MPH on NIRS signals after single-dose, 4-to-8-week, and 1-year administration. ΔNIRS significantly predicted CGI-S after 4-to-8-week MPH administration. The leave-one-out classification algorithm had 81% accuracy using the NIRS signal. ΔNIRS also significantly predicted CGI-S scores after 1 year of MPH administration. For secondary analyses, NAÏVE exhibited significantly lower prefrontal activation than HCs at the baseline assessment, whereas NON-NAÏVE and HCs showed similar activation. A single dose of MPH significantly increased activation compared with the placebo in NAÏVE. After 4-to-8-week administration, and even after MPH washout following 1-year administration, NAÏVE demonstrated normalized prefrontal activation. Supplementary NIRS measurements may serve as an objective biomarker for clinical decisions and monitoring concerning continuous MPH treatment in children with ADHD.

Genetic influences on prefrontal activation during a verbal fluency task in adults: a twin study based on multichannel near-infrared spectroscopy.

Near-infrared spectroscopy (NIRS) studies have reported that prefrontal hemodynamic dysfunction during executive function tasks may be a promising biomarker of psychiatric disorders, because its portability and noninvasiveness allow easy measurements in clinical settings. Here, we investigated the degree to which prefrontal NIRS signals are genetically determined. Using a 52-channel NIRS system, we monitored the oxy-hemoglobin (oxy-Hb) signal changes in 38 adult pairs of right-handed monozygotic (MZ) twins and 13 pairs of same-sex right-handed dizygotic (DZ) twins during a letter version of the verbal fluency task. Heritability was estimated based on a classical twin paradigm using structured equation modeling. Significant genetic influences were estimated in the right dorsolateral prefrontal cortex and left frontal pole. The degrees of heritability were 66% and 75% in the variances, respectively. This implies that the prefrontal hemodynamic dysfunction observed during an executive function task measured by NIRS may be an efficient endophenotype for large-scale imaging genetic studies in psychiatric disorders.

Prefrontal activation during inhibitory control measured by near-infrared spectroscopy for differentiating between autism spectrum disorders and attention deficit hyperactivity disorder in adults.

The differential diagnosis of autism spectrum disorders (ASDs) and attention deficit hyperactivity disorder (ADHD) based solely on symptomatic and behavioral assessments can be difficult, even for experts. Thus, the development of a neuroimaging marker that differentiates ASDs from ADHD would be an important contribution to this field. We assessed the differences in prefrontal activation between adults with ASDs and ADHD using an entirely non-invasive and portable neuroimaging tool, near-infrared spectroscopy. This study included 21 drug-naïve adults with ASDs, 19 drug-naïve adults with ADHD, and 21 healthy subjects matched for age, sex, and IQ. Oxygenated hemoglobin concentration changes in the prefrontal cortex were assessed during a stop signal task and a verbal fluency task. During the stop signal task, compared to the control group, the ASDs group exhibited lower activation in a broad prefrontal area, whereas the ADHD group showed underactivation of the right premotor area, right presupplementary motor area, and bilateral dorsolateral prefrontal cortices. Significant differences were observed in the left ventrolateral prefrontal cortex between the ASDs and ADHD groups during the stop signal task. The leave-one-out cross-validation method using mean oxygenated hemoglobin changes yielded a classification accuracy of 81.4% during inhibitory control. These results were task specific, as the brain activation pattern observed during the verbal fluency task did not differentiate the ASDs and ADHD groups significantly. This study therefore provides evidence of a difference in left ventrolateral prefrontal activation during inhibitory control between adults with ASDs and ADHD. Thus, near-infrared spectroscopy may be useful as an auxiliary tool for the differential diagnosis of such developmental disorders.

Parietal EEG alpha suppression time of memory retrieval reflects memory load while the alpha power of memory maintenance is a composite of the visual process according to simultaneous and successive Sternberg memory tasks.

The present study investigated EEG alpha activity during visual Sternberg memory tasks using two different stimulus presentation modes to elucidate how the presentation mode affected parietal alpha activity. EEGs were recorded from 10 healthy adults during the Sternberg tasks in which memory items were presented simultaneously and successively. EEG power and suppression time (ST) in the alpha band (8-13Hz) were computed for the memory maintenance and retrieval phases. The alpha activity differed according to the presentation mode during the maintenance phase but not during the retrieval phase. Results indicated that parietal alpha power recorded during the maintenance phase did not reflect the memory load alone. In contrast, ST during the retrieval phase increased with the memory load for both presentation modes, indicating a serial memory scanning process, regardless of the presentation mode. These results indicate that there was a dynamic transition in the memory process from the maintenance phase, which was sensitive to external factors, toward the retrieval phase, during which the process converged on the sequential scanning process, the Sternberg task essentially required.

EEG coherence pattern during simultaneous and successive processing tasks.

In order to investigate psychophysiological basis of two different types of information processing (simultaneous and successive), we investigated EEG coherence patterns during five tasks of the Das-Naglieri Cognitive Assessment System. The investigation was based on the perspective of the similarities among the tasks of the same processing type and discrepancies between the different processing types. EEG coherence of eighteen volunteers during two simultaneous and three successive tasks varying in task content (verbal-nonverbal) and modality (auditory-visual) was computed. First, we compared the coherence patterns between the tasks of the same processing type, then compared between the tasks of different processing types. Result revealed a specific theta coherence pattern associated with processing type regardless of task contents and modality. Theta coherence during the simultaneous processing showed increased short-range inter-hemispheric connections over central and parietal regions as compared to the successive tasks. Upper beta coherence indicated task-dependent effect regardless of processing types. In addition, our data indicated no hemispheric differences for both processing types. It suggested that confounding of verbal-nonverbal and processing type (simultaneous-successive) dimension could be avoided. Our result presented psychophysiological evidence for the existence of two types of information processing, which has been supported by psychological studies with factor analysis.

Differential topographic pattern of EEG coherence between simultaneous and successive coding tasks.

The concept of two types of information coding, simultaneous and successive processing, is now well supported by extensive studies with factor analysis. However, few EEG evidence on processing types have been reported. In the present study we investigated whether varying demands on simultaneous or successive processing are reflected by different pattern of EEG coherence change from the passive condition to the active condition. We computed EEG coherence during simultaneous and successive processing tasks in both passive and active conditions. Under the passive condition, participants were just to perceive the presented stimuli. In the active condition, participants were required to remember the presented stimuli and then reproduce or recognize the remembered stimuli. Our result revealed the different topographic patterns of coherence change from the passive to the active condition between the simultaneous and the successive task. In the successive processing task, bilateral frontal-left temporal coherence in beta showed a significant decrease during the active condition, supporting Luria's model of the two information coding types. The condition effect of coherence in the simultaneous processing task was rather unclear. Our data also indicated that more task related cognitive processes, rather than the task-independent processes such as attentional demand, were reflected in EEG coherence of higher frequency bands. The different EEG coherence patterns seen in the simultaneous and successive tasks suggested the first step evidence that EEG coherence pattern may differentiate two distinctive types of information coding.