Recent studies have suggested oxytocin’s therapeutic effects on deficits in social communication and interaction in autism spectrum disorder through improvement of emotion recognition with direct emotional cues, such as facial expression and voice prosody. Although difficulty in understanding of others’ social emotions and beliefs under conditions without direct emotional cues also plays an important role in autism spectrum disorder, no study has examined the potential effect of oxytocin on this difficulty. Here, we sequentially conducted both a case-control study and a clinical trial to investigate the potential effects of oxytocin on this difficulty at behavioural and neural levels measured using functional magnetic resonance imaging during a psychological task. This task was modified from the Sally-Anne Task, a well-known first-order false belief task. The task was optimized for investigation of the abilities to infer another person’s social emotions and beliefs distinctively so as to test the hypothesis that oxytocin improves deficit in inferring others’ social emotions rather than beliefs, under conditions without direct emotional cues. In the case-control study, 17 males with autism spectrum disorder showed significant behavioural deficits in inferring others’ social emotions (P = 0.018) but not in inferring others’ beliefs (P = 0.064) compared with 17 typically developing demographically-matched male participants. They also showed significantly less activity in the right anterior insula and posterior superior temporal sulcus during inferring others’ social emotions, and in the dorsomedial prefrontal cortex during inferring others’ beliefs compared with the typically developing participants (P < 0.001 and cluster size > 10 voxels). Then, to investigate potential effects of oxytocin on these behavioural and neural deficits, we conducted a double-blind placebo-controlled crossover within-subject trial for single-dose intranasal administration of 24 IU oxytocin in an independent group of 20 males with autism spectrum disorder. Behaviourally, oxytocin significantly increased the correct rate in inferring others’ social emotions (P = 0.043, one-tail). At the neural level, the peptide significantly enhanced the originally-diminished brain activity in the right anterior insula during inferring others’ social emotions (P = 0.004), but not in the dorsomedial prefrontal cortex during inferring others’ beliefs (P = 0.858). The present findings suggest that oxytocin enhances the ability to understand others’ social emotions that have also required second-order false belief rather than first-order false beliefs under conditions without direct emotional cues in autism spectrum disorder at both the behaviour and neural levels.

Neuropsychiatric disorders are diagnosed based on behavioral criteria, which makes the diagnosis challenging. Objective biomarkers such as neuroimaging are needed, and when coupled with machine learning, can assist the diagnostic decision and increase its reliability. Sixty-four schizophrenia, 36 autism spectrum disorder (ASD), and 106 typically developing individuals were analyzed. FreeSurfer was used to obtain the data from the participant’s brain scans. Six classifiers were utilized to classify the subjects. Subsequently, 26 ultra-high risk for psychosis (UHR) and 17 first-episode psychosis (FEP) subjects were run through the trained classifiers. Lastly, the classifiers’ output of the patient groups was correlated with their clinical severity. All six classifiers performed relatively well to distinguish the subject groups, especially support vector machine (SVM) and Logistic regression (LR). Cortical thickness and subcortical volume feature groups were most useful for the classification. LR and SVM were highly consistent with clinical indices of ASD. When UHR and FEP groups were run with the trained classifiers, majority of the cases were classified as schizophrenia, none as ASD. Overall, SVM and LR were the best performing classifiers. Cortical thickness and subcortical volume were most useful for the classification, compared to surface area. LR, SVM, and DT’s output were clinically informative. The trained classifiers were able to help predict the diagnostic category of both UHR and FEP Individuals.

Purpose: Large language models (LLMs) are rapidly advancing and demonstrating high performance in understanding textual information, suggesting potential applications in interpreting patient histories and documented imaging findings. As LLMs continue to improve, their diagnostic abilities are expected to be enhanced further. However, there is a lack of comprehensive comparisons between LLMs from different manufacturers. In this study, we aimed to test the diagnostic performance of the three latest major LLMs (GPT-4o, Claude 3 Opus, and Gemini 1.5 Pro) using Radiology Diagnosis Please Cases, a monthly diagnostic quiz series for radiology experts. Materials and methods: Clinical history and imaging findings, provided textually by the case submitters, were extracted from 324 quiz questions originating from Radiology Diagnosis Please cases published between 1998 and 2023. The top three differential diagnoses were generated by GPT-4o, Claude 3 Opus, and Gemini 1.5 Pro, using their respective application programming interfaces. A comparative analysis of diagnostic performance among these three LLMs was conducted using Cochrane’s Q and post hoc McNemar’s tests. Results: The respective diagnostic accuracies of GPT-4o, Claude 3 Opus, and Gemini 1.5 Pro for primary diagnosis were 41.0%, 54.0%, and 33.9%, which further improved to 49.4%, 62.0%, and 41.0%, when considering the accuracy of any of the top three differential diagnoses. Significant differences in the diagnostic performance were observed among all pairs of models. Conclusion: Claude 3 Opus outperformed GPT-4o and Gemini 1.5 Pro in solving radiology quiz cases. These models appear capable of assisting radiologists when supplied with accurate evaluations and worded descriptions of imaging findings.

Changes in brain pathology as schizophrenia progresses have been repeatedly suggested by previous studies. Meta-analyses of previous proton magnetic resonance spectroscopy (¹H MRS) studies at each clinical stage of schizophrenia indicate that the abnormalities of N-acetylaspartate (NAA) and glutamatergic metabolites change progressively. However, to our knowledge, no single study has addressed the possible differences in ¹H MRS abnormalities in subjects at 3 different stages of disease, including those at ultrahigh risk for psychosis (UHR), with first-episode schizophrenia (FES), and with chronic schizophrenia (ChSz). In the current study, 24 patients with UHR, 19 FES, 25 ChSz, and their demographically matched 3 independent control groups (n = 26/19/28 for the UHR, FES, and ChSz control groups, respectively) underwent ¹H MRS in a 3-Tesla scanner to examine metabolites in medial prefrontal cortex. The analysis revealed significant decreases in the medial prefrontal NAA and glutamate + glutamine (Glx) levels, specifically in the ChSz group as indexed by a significant interaction between stage (UHR/FES/ChSz) and clinical status (patients/controls) (P = .008). Furthermore, the specificity of NAA and Glx reductions compared with the other metabolites in the patients with ChSz was also supported by a significant interaction between the clinical status and types of metabolites that only occurred at the ChSz stage (P = .001 for NAA, P = .004 for Glx). The present study demonstrates significant differences in ¹H MRS abnormalities at different stages of schizophrenia, which potentially correspond to changes in glutamatergic neurotransmission, plasticity, and/or excitotoxicity and regional neuronal integrity with relevance for the progression of schizophrenia.

Purpose: The diagnostic performance of large language artificial intelligence (AI) models when utilizing radiological images has yet to be investigated. We employed Claude 3 Opus (released on March 4, 2024) and Claude 3.5 Sonnet (released on June 21, 2024) to investigate their diagnostic performances in response to the Radiology’s Diagnosis Please quiz questions. Materials and methods: In this study, the AI models were tasked with listing the primary diagnosis and two differential diagnoses for 322 quiz questions from Radiology’s “Diagnosis Please” cases, which included cases 1 to 322, published from 1998 to 2023. The analyses were performed under the following conditions: (1) Condition 1: submitter-provided clinical history (text) alone. (2) Condition 2: submitter-provided clinical history and imaging findings (text). (3) Condition 3: clinical history (text) and key images (PNG file). We applied McNemar’s test to evaluate differences in the correct response rates for the overall accuracy under Conditions 1, 2, and 3 for each model and between the models. Results: The correct diagnosis rates were 58/322 (18.0%) and 69/322 (21.4%), 201/322 (62.4%) and 209/322 (64.9%), and 80/322 (24.8%) and 97/322 (30.1%) for Conditions 1, 2, and 3 for Claude 3 Opus and Claude 3.5 Sonnet, respectively. The models provided the correct answer as a differential diagnosis in up to 26/322 (8.1%) for Opus and 23/322 (7.1%) for Sonnet. Statistically significant differences were observed in the correct response rates among all combinations of Conditions 1, 2, and 3 for each model (p < 0.01). Claude 3.5 Sonnet outperformed in all conditions, but a statistically significant difference was observed only in the comparison for Condition 3 (30.1% vs. 24.8%, p = 0.028). Conclusion: Two AI models demonstrated a significantly improved diagnostic performance when inputting both key images and clinical history. The models’ ability to identify important differential diagnoses under these conditions was also confirmed.