1,720,999 research outputs found
AesFA: An Aesthetic Feature-Aware Arbitrary Neural Style Transfer
No full textNeural style transfer (NST) has evolved significantly in recent years. Yet, despite its rapid progress and advancement, existing NST methods either struggle to transfer aesthetic information from a style effectively or suffer from high computational costs and inefficiencies in feature disentanglement due to using pre-trained models. This work proposes a lightweight but effective model, AesFA-Aesthetic Feature-Aware NST. The primary idea is to decompose the image via its frequencies to better disentangle aesthetic styles from the reference image while training the entire model in an end-to-end manner to exclude pre-trained models at inference completely. To improve the network's ability to extract more distinct representations and further enhance the stylization quality, this work introduces a new aesthetic feature: contrastive loss. Extensive experiments and ablations show the approach not only outperforms recent NST methods in terms of stylization quality, but it also achieves faster inference. Codes are available at https://github.com/Sooyyoungg/AesFA.Y
3D distributed deep learning framework for prediction of human intelligence from brain MRI
No full textIntelligence is a complex, multi-dimensional concept that encompasses multiple brain circuits. Understanding the underpinnings of the human brain requires not only accurate feature extraction from often noisy non-invasive brain imaging data (e.g., MRI), but also rigorous modeling of the complex relationships among distributed brain systems. In this work, we implement a highly scalable end-to-end computational learning framework - that is, a 3D deep convolutional neural network (CNN) to predict fluid intelligence scores directly from 3D brain MRI without any theory- or rule-based feature engineering. We address and overcome the challenge of processing large data (i.e. 44 GB of MRI) by using distributed deep learning techniques. The dataset originates from the Adolescent Brain Cognitive Development (ABCD) study, with 5832 subjects in the training set, 1251 in the validation set, and 1250 in the test set. The single-task ResNet50-3D model achieved mean squared errors of 0.73637 and 0.74535 respectively on the validation and test sets. The multi-task ResNet50-3D model achieved mean squared errors of 0.74418 and 0.75626 respectively on the validation and test sets. These results demonstrate not only that the prediction of fluid intelligence scores directly from structural and diffusion brain MRI is feasible but also that this scalable computational learning framework could be further developed for data-driven human neurocognitive research.N
Therapeutic Effect of Repetitive Transcranial Magnetic Stimulation for Post-stroke Vascular Cognitive Impairment: A Prospective Pilot Study
No full textObjectivePost-stroke cognitive impairment (PSCI) is resistant to treatment. Recent studies have widely applied repetitive transcranial magnetic stimulation (rTMS) to treat various brain dysfunctions, such as post-stroke syndromes. Nonetheless, a protocol for PSCI has not been established. Therefore, this study is aimed to evaluate the therapeutic effect of our high-frequency rTMS protocol for PSCI during the chronic phase of stroke. MethodsIn this prospective study, ten patients with PSCI were enrolled and received high-frequency rTMS on the ipsilesional dorsolateral prefrontal cortex (DLPFC) for 10 sessions (5 days per week for 2 weeks). Cognitive and affective abilities were assessed at baseline and 2 and 14 weeks after rTMS initiation. To investigate the therapeutic mechanism of rTMS, the mRNA levels of pro-inflammatory cytokines (interleukin (IL)-6, IL-1 beta, transforming growth factor beta [TGF-beta], and tumor necrosis factor alpha [TNF-alpha]) in peripheral blood samples were quantified using reverse transcription polymerase chain reaction, and cognitive functional magnetic resonance imaging (fMRI) was conducted at baseline and 14 weeks in two randomly selected patients after rTMS treatment. ResultsThe scores of several cognitive evaluations, i.e., the Intelligence Quotient (IQ) of Wechsler Adult Intelligence Scale, auditory verbal learning test (AVLT), and complex figure copy test (CFT), were increased after completion of the rTMS session. After 3 months, these improvements were sustained, and scores on the Mini-Mental Status Examination and Montreal Cognitive Assessment (MoCA) were also increased (p < 0.05). While the Geriatric Depression Scale (GeDS) did not show change among all patients, those with moderate-to-severe depression showed amelioration of the score, with marginal significance. Expression of pro-inflammatory cytokines was decreased immediately after the ten treatment sessions, among which, IL-1 beta remained at a lower level after 3 months. Furthermore, strong correlations between the decrease in IL-6 and increments in AVLT (r = 0.928) and CFT (r = 0.886) were found immediately after the rTMS treatment (p < 0.05). Follow-up fMRI revealed significant activation in several brain regions, such as the medial frontal lobe, hippocampus, and angular area. ConclusionsHigh-frequency rTMS on the ipsilesional DLPFC may exert immediate efficacy on cognition with the anti-inflammatory response and changes in brain network in PSCI, lasting at least 3 months.N
Diagnosis and prognosis of Alzheimer's disease using brain morphometry and white matter connectomes
No full textAccurate, reliable prediction of risk for Alzheimer's disease (AD) is essential for early, disease-modifying therapeutics. Multimodal MRI, such as structural and diffusion MRI, is likely to contain complementary information of neurodegenerative processes in AD. Here we tested the utility of the multimodal MRI (T1-weighted structure and diffusion MRI), combined with high-throughput brain phenotyping-morphometry and structural connectomics-and machine learning, as a diagnostic tool for AD. We used, firstly, a clinical cohort at a dementia clinic (National Health Insurance Service-Ilsan Hospital [NHIS-IH]; N = 211; 110 AD, 64 mild cognitive impairment [MCI], and 37 cognitively normal with subjective memory complaints [SMC]) to test the diagnostic models; and, secondly, Alzheimer's Disease Neuroimaging Initiative (ADNI)-2 to test the generalizability. Our machine learning models trained on the morphometric and connectome estimates (number of features = 34,646) showed optimal classification accuracy (AD/SMC: 97% accuracy, MCI/SMC: 83% accuracy; AD/ MCI: 97% accuracy) in NHIS-IH cohort, outperforming a benchmark model (FLAIR-based white matter hyperintensity volumes). In ADNI-2 data, the combined connectome and morphometry model showed similar or superior accuracies (AD/HC: 96%; MCI/HC: 70%; AD/MCI: 75% accuracy) compared with the CSF biomarker model (t-tau, p-tau, and Amyloid beta, and ratios). In predicting MCI to AD progression in a smaller cohort of ADNI-2 (n = 60), the morphometry model showed similar performance with 69% accuracy compared with CSF biomarker model with 70% accuracy. Our comparisons of the classifiers trained on structural MRI, diffusion MRI, FLAIR, and CSF biomarkers showed the promising utility of the white matter structural connectomes in classifying AD and MCI in addition to the widely used structural MRI-based morphometry, when combined with machine learning.Y
Cognitive Capacity Genome-Wide Polygenic Scores Identify Individuals with Slower Cognitive Decline in Aging
No full textThe genetic protective factors for cognitive decline in aging remain unknown. Predicting an individual’s rate of cognitive decline—or with better cognitive resilience—using genetics will allow personalized intervention for cognitive enhancement and the optimal selection of target samples in clinical trials. Here, using genome-wide polygenic scores (GPS) of cognitive capacity as the genomic indicators for variations of human intelligence, we analyzed the 18-year records of cognitive and behavioral data of 8511 European-ancestry adults from the Wisconsin Longitudinal Study (WLS), specifically focusing on the cognitive assessments that were repeatedly administered to the participants with their average ages of 64.5 and 71.5. We identified a significant interaction effect between age and cognitive capacity GPS, which indicated that a higher cognitive capacity GPS significantly correlated with a slower cognitive decline in the domain of immediate memory recall (β = 1.86 × 10(−1), p-value = 1.79 × 10(−3)). The additional phenome-wide analyses identified several associations between cognitive capacity GPSs and cognitive/behavioral phenotypes, such as similarities task (β = 1.36, 95% CI = (1.22, 1.51), p-value = 3.59 × 10(−74)), number series task (β = 0.94, 95% CI = (0.85, 1.04), p-value = 2.55 × 10(−78)), IQ scores (β = 1.42, 95% CI = (1.32, 1.51), p-value = 7.74 × 10(−179)), high school classrank (β = 1.86, 95% CI = (1.69, 2.02), p-value = 3.07 × 10(−101)), Openness from the BIG 5 personality factor (p-value = 2.19 × 10(−14), β = 0.57, 95% CI = (0.42, 0.71)), and leisure activity of reading books (β = 0.50, 95% CI = (0.40, 0.60), p-value = 2.03 × 10(−21)), attending cultural events, such as concerts, plays, or museums (β = 0.60, 95% CI = (0.49, 0.72), p-value = 2.06 × 10(−23)), and watching TV (β = −0.48, 95% CI = (−0.59, −0.37), p-value = 4.16 × 10(−18)). As the first phenome-wide analysis of cognitive and behavioral phenotypes, this study presents the novel genetic protective effects of cognitive ability on the decline of memory recall in an aging population
SwiFT: Swin 4D fMRI Transformer
No full textModeling spatiotemporal brain dynamics from high-dimensional data, such as functional Magnetic Resonance Imaging (fMRI), is a formidable task in neuroscience. Existing approaches for fMRI analysis utilize hand-crafted features, but the process of feature extraction risks losing essential information in fMRI scans. To address this challenge, we present SwiFT (Swin 4D fMRI Transformer), a Swin Transformer architecture that can learn brain dynamics directly from fMRI volumes in a memory and computation-efficient manner. SwiFT achieves this by implementing a 4D window multi-head self-attention mechanism and absolute positional embeddings. We evaluate SwiFT using multiple large-scale resting-state fMRI datasets, including the Human Connectome Project (HCP), Adolescent Brain Cognitive Development (ABCD), and UK Biobank (UKB) datasets, to predict sex, age, and cognitive intelligence. Our experimental outcomes reveal that SwiFT consistently outperforms recent state-of-the-art models. Furthermore, by leveraging its end-to-end learning capability, we show that contrastive loss-based self-supervised pre-training of SwiFT can enhance performance on downstream tasks. Additionally, we employ an explainable AI method to identify the brain regions associated with sex classification. To our knowledge, SwiFT is the first Swin Transformer architecture to process dimensional spatiotemporal brain functional data in an end-to-end fashion. Our work holds substantial potential in facilitating scalable learning of functional brain imaging in neuroscience research by reducing the hurdles associated with applying Transformer models to high-dimensional fMRI. Project page: https://github.com/Transconnectome/SwiFTN
Early life stress modulates the genetic influence on brain structure and cognitive function in children
No full textThe enduring influence of early life stress (ELS) on brain and cognitive development has been widely acknowledged, yet the precise mechanisms underlying this association remain elusive. We hypothesize that ELS might disrupt the genome-wide influence on brain morphology and connectivity development, consequently exerting a detrimental impact on children's cognitive ability. We analyzed the multimodal data of DNA genotypes, brain imaging (structural and diffusion MRI), and neurocognitive battery (NIH Toolbox) of 4276 children (ages 9-10 years, European ancestry) from the Adolescent Brain Cognitive Development (ABCD) study. The genome-wide influence on cognitive function was estimated using the polygenic score (GPS). By using brain morphometry and tractography, we identified the brain correlates of the cognition GPSs. Statis-tical analyses revealed relationships for the gene-brain-cognition pathway. The brain structural variance significantly mediated the genetic influence on cognition (indirect effect = 0.016, P-FDR < 0.001). Of note, this gene-brain relationship was significantly modulated by abuse, resulting in diminished cognitive capacity (Index of Moderated Mediation = -0.007; 95 % CI = -0.012 similar to -0.002). Our results support a novel gene-brain-cognition model likely elucidating the longlasting negative impact of ELS on children's cognitive development.Y
Machine learning prediction of incidence of Alzheimer’s disease using large-scale administrative health data
Full text linkNationwide population-based cohort provides a new opportunity to build an automated risk prediction model based on individuals’ history of health and healthcare beyond existing risk prediction models. We tested the possibility of machine learning models to predict future incidence of Alzheimer’s disease (AD) using large-scale administrative health data. From the Korean National Health Insurance Service database between 2002 and 2010, we obtained de-identified health data in elders above 65 years (N = 40,736) containing 4,894 unique clinical features including ICD-10 codes, medication codes, laboratory values, history of personal and family illness and socio-demographics. To define incident AD we considered two operational definitions: “definite AD” with diagnostic codes and dementia medication (n = 614) and “probable AD” with only diagnosis (n = 2026). We trained and validated random forest, support vector machine and logistic regression to predict incident AD in 1, 2, 3, and 4 subsequent years. For predicting future incidence of AD in balanced samples (bootstrapping), the machine learning models showed reasonable performance in 1-year prediction with AUC of 0.775 and 0.759, based on “definite AD” and “probable AD” outcomes, respectively; in 2-year, 0.730 and 0.693; in 3-year, 0.677 and 0.644; in 4-year, 0.725 and 0.683. The results were similar when the entire (unbalanced) samples were used. Important clinical features selected in logistic regression included hemoglobin level, age and urine protein level. This study may shed a light on the utility of the data-driven machine learning model based on large-scale administrative health data in AD risk prediction, which may enable better selection of individuals at risk for AD in clinical trials or early detection in clinical settings
Identifying Prepubertal Children with Risk for Suicide Using Deep Neural Network Trained on Multimodal Brain Imaging
No full text© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.Suicide is among the leading causes of death in youth worldwide. Early identification of children with high risk for suicide is a key to effective screening and intervention strategies. Yet, little is known about the neural pathways to the clinical outcomes of youth suicide. In this study, we tested brain functional substrates associated with the risk for youth suicidality. Based on the large, multi-site, multi-ethnic, representative, and prospective developmental population study in the US, we trained a state-of-the-art interpretable deep neural network on functional brain imaging, behavioral, and self-reported questionnaires. Our best model contains the functional estimates of key brain regions important for attention, emotion regulation, and motor coordination, such as the anterior cingulate cortex, temporal gyrus, and precentral gyrus. The interpretable neural network shows that these brain functional features interact with depression and impulsivity, the known risk factors of youth suicidality. This study demonstrates a novel application of the interpretable deep neural network to childhood suicidal research, uncovering the complex interactions between psychological and neural factors underlying youth suicidality.N
Maturity of gray matter structures and white matter connectomes, and their relationship with psychiatric symptoms in youth
No full textBrain predicted age difference, or BrainPAD, compares chronological age to an age estimate derived by applying machine learning (ML) to MRI brain data. BrainPAD studies in youth have been relatively limited, often using only a single MRI modality or a single ML algorithm. Here, we use multimodal MRI with a stacked ensemble ML approach that iteratively applies several ML algorithms (AutoML). Eligible participants in the Healthy Brain Network (N = 489) were split into training and test sets. Morphometry estimates, white matter connectomes, or both were entered into AutoML to develop BrainPAD models. The best model was then applied to a held-out evaluation dataset, and associations with psychometrics were estimated. Models using morphometry and connectomes together had a mean absolute error of 1.18 years, outperforming models using a single MRI modality. Lower BrainPAD values were associated with more symptoms on the CBCL (p(corr) = .012) and lower functioning on the Children's Global Assessment Scale (p(corr) = .012). Higher BrainPAD values were associated with better performance on the Flanker task (p(corr) = .008). Brain age prediction was more accurate using ComBat-harmonized brain data (MAE = 0.26). Associations with psychometric measures remained consistent after ComBat harmonization, though only the association with CGAS reached statistical significance in the reduced sample. Our findings suggest that BrainPAD scores derived from unharmonized multimodal MRI data using an ensemble ML approach may offer a clinically relevant indicator of psychiatric and cognitive functioning in youth.Y
- …
