Biology subjects

Lin, F. V.

Publications and source records attributed to Lin, F. V..

8 recordsLinked to original sources

bioRxiv2026

Benchmarking foundation models for resting-state functional magnetic resonance imaging time series in cognitive aging research

Foundation models trained on large neuroimaging datasets could help smaller studies investigate cognition from brain recordings. However, whether the foundation model-derived representations improve assessment of current cognition, prediction of follow-up cognition, or characterization of cognitive change remains unclear. We evaluated nine configurations from five model families that process regional resting-state functional magnetic resonance imaging time series, using 3,103 scans from 1,653 participants across two observational and five intervention aging cohorts. Primary frozen representations showed no consistent superiority over conventional functional connectivity for concurrent cognitive associations or diagnostic classification. Changing the readout improved selected concurrent cognitive associations without updating model weights, indicating that performance depended partly on how representations were extracted. Additional post-training and embedding alignment yielded gains in selected settings but did not consistently improve performance across outcomes. In the primary longitudinal analyses, imaging representations did not establish added value beyond baseline clinical measures. These findings show that readout choice influences concurrent cognitive associations, while the added value of these representations for longitudinal outcomes remains unestablished.

bioengineering↗

bioRxiv2026

Bilingualism Protects Domain-Specific Cognitive Function in Mandarin-Speaking Older Adults

ObjectivesThis study examined whether bilingualism is associated with episodic memory (EM) and executive function (EF) in older Mandarin-speaking adults and whether associations differ by clinical diagnosis. Methods189 Mandarin-speaking older adults completed Mandarin-administered neuropsychological testing, brain MRI, and clinical diagnosis. English proficiency test was administered to determine whether they were bilinguals or monolinguals. Sensitivity analyses were conducted in an education-matched subgroup (16 years). ResultsBilingualism was associated with higher EF, with a significant bilingualism x diagnosis interaction indicating larger bilingual advantages among cognitively impaired participants. However, bilingualism was not associated with EM. Findings were preserved in education-matched analysis. ConclusionsResults support a domain-specific association between bilingualism and executive function in later life, consistent with cognitive maintenance mechanisms preferentially supporting executive processes rather than global protection.

neuroscience↗

bioRxiv2025

Medial Temporal Default Mode Network Selectively Encodes Autobiographical Visual Imagery

The human brains capacity to imagine visual scenes from memory is thought to rely on the medial temporal subsystem of the default mode network (MT-DMN), yet the neural codes supporting this ability remain poorly understood. We combined functional magnetic resonance imaging (fMRI) with vision and language artificial intelligence models to characterize neural codes during autobiographical imagination. Fifty participants imagined reexperiencing twenty natural scenarios while undergoing fMRI, when cued by generic text prompts (e.g., wedding, exercising, driving). Individual scenes were modeled using Stable Diffusion to generate personalized synthetic images from verbal descriptions of the scenarios imagined, collected beforehand. These depictions were then transformed into image-recognition network embeddings. Representational Similarity Analysis revealed that the MT-DMN encoded the participant-specific representational structure of image embeddings, even when controlling for semantic features derived from a large language model. This effect was absent in other networks and during reading without imagination, identifying the MT-DMN as a core substrate for the visual reconstruction of autobiographical experiences.

neuroscience↗

bioRxiv2025

A default-control network cortical gradient differentiates the imagination of social and solitary experiences

Understanding the neural basis of spontaneous thought - when attention shifts from external tasks to internally generated content such as reminiscing or planning - remains a central challenge in cognitive neuroscience. Progress in this area has lagged studies of externally driven cognition, in part because self-generated thought is difficult to control and measure. Recent work suggests that transitions between externally and internally focused cognitive states follow a continuous neural activation gradient that reflects opposing patterns of engagement of the frontoparietal control network versus the default mode network. To characterize internally focused cognitive states that prospectively engage the default mode network, we used functional Magnetic Resonance Imaging (fMRI) to measure brain activity while participants imagined a range of personal scenarios prompted by generic text cues (e.g., party, housework) to mimic natural thought. Before scanning, participants described each scenario verbally and rated its experiential feature content. Gradient-space analysis revealed a cognitive transition within imagined states: solitary activities (e.g., housework) preferentially recruited the frontoparietal network relative to the default mode network, whereas social activities (e.g., a party) showed the opposite pattern. Because all cognitive states were internally focused (imagined) and elicited via the same non-interactive task, these results refine interpretations of the frontoparietal-default mode gradient. They show that this gradient does not simply differentiate external task-positive from internal task-negative states but also tracks semantic differences between internal cognitive states.

neuroscience↗

bioRxiv2025

AI-assisted modeling of attentional control for intervention engagement

Digital cognitive interventions can improve cognition in older adults, but individual differences in transfer to executive function (EF) remain substantial and are difficult to anticipate before treatment completion. We developed a compact Facial Signature Score (FSS) from passive webcam recordings collected during therapeutic cognitive challenges in BREATHE (NCT04522791; development, N=87) and evaluated it in PSOPT (NCT06005038; external validation, N=39), an independent trial using a different intervention format. The FSS discriminated EF responders under nested cross-validation in BREATHE (AUROC=0.77) and generalized to PSOPT without feature reselection or reweighting (AUROC=0.78). The FSS was associated with continuous EF change in BREATHE (Spearman {rho}=0.49, p<0.001) and after outcome-independent recalibration in PSOPT ({rho}=0.32, p=0.045). Predictive performance was retained when the signature was computed from the first 25% and 50% of intervention sessions, and the FSS added value beyond baseline characteristics and conventional behavioral measures. Passive facial dynamics during therapeutic cognitive challenge may serve as a scalable intervention-process marker for early assessment of cognitive response among older adults.

neuroscience↗

bioRxiv2025

Functional Connectome of Superagers Reveals Early Markers of Resilience and Vulnerability to Alzheimers Disease

As populations age, identifying the neurobiological basis of cognitive resilience is critical for delaying or preventing Alzheimers disease (AD). While most older adults experience memory decline, a subset known as superagers (SA) maintains youthful memory into late life, offering a unique window into protective mechanisms against neurodegeneration. Here, we identified a functional connectivity (FC) signature, termed Alzheimers-resilient connectome (ARC), that robustly differentiates SA from age-matched patients with AD. Using resting-state fMRI in a discovery cohort (N = 290), we identified ARC derived from machine learning classifiers that distinguished SA from AD with high accuracy (AUC = 0.85), and validated the replicability of the ARC in an independent replication cohort (N = 143). ARC involved prefrontal, temporal and insular networks and was strongly associated with brain age. When applied to cognitively unimpaired (CU) adults (discovery cohort: N = 818 and replication cohort: N = 497), ARC-based subtyping revealed SA-like and AD-like subgroups with similar baseline cognitive performance but markedly divergent longitudinal trajectories. SA-like CU individuals showed slower cognitive decline, reduced amyloid-{beta} accumulation, and lower risk of conversion to mild cognitive impairment and AD, reinforcing the ARC signature as a potential early indicator of resilience. Genome-wide association analysis identified CLYBL and FRMD6 as novel genetic modulators associated with these divergent aging phenotypes. Together, our findings position ARC as a sensitive and generalizable biomarker of resilience, enabling early risk stratification and precision prevention for AD.

neuroscience↗

bioRxiv2023

The impact of multiband and in-plane acceleration on white matter microstructure analysis

Accelerated imaging has been broadly adopted in diffusion MRI studies, yet little is known about its impacts. Acceleration can achieve higher spatial and q-space resolution in shorter time, reduce motion artifacts, and reduce patient burden. However, it leads to noise amplification, and its impacts in clinical cohorts are poorly understood. This study examined the impact of multiband (also called simultaneous multislice, or SMS) and in-plane acceleration (IPA, also called phase acceleration) in diffusion imaging in forty older adults differing in cognitive status. We evaluated a total of 400 scans from five acquisitions: no acceleration (S1P1); SMS=3 with no in-plane acceleration (S3P1); SMS=3 with IPA=2 (S3P2); S6P1; and S6P2. The number of diffusion directions and b-values was kept constant such that acquisition times varied from 21:28 to 3:56. We found that diffusion metrics were highly sensitive to acceleration factor, with a trend towards higher fractional anisotropy (FA) and lower orientation dispersion (OD) with acceleration. The differences between accelerated and unaccelerated acquisitions could be partly explained by the noise amplification (g-factor) and reduced motion. Intraclass correlations (ICCs) of FA and OD in white matter were excellent in both S1P1 and S3P1 (median >0.8), good but lower in S3P2 and S6P1 (medians around 0.70), and poor to fair in S6P2 (medians 0.46 and 0.57). In-plane acceleration decreased ICC, including areas of high susceptibility distortion. In a comparison of mild cognitive impairment versus healthy controls, acceleration tended to reduce group differences, particularly in the fornix, with greater costs in OD than FA. Our results provide guidance regarding the costs of acceleration (possible biases and reduced effect sizes) while also characterizing the benefits (reduced motion, good reliability at higher multiband with no in-plane).

bioengineering↗

bioRxiv2020

Learning Clique Subgraphs in Structural Brain Network Classification with Application to Crystallized Cognition

Structural brain networks constructed from diffusion MRI are important biomarkers for understanding human brain structure and its relation to cognitive functioning. There is increasing interest in learning differences in structural brain networks between groups of subjects in neuroimaging studies, leading to a variable selection problem in network classification. Traditional methods often use independent edgewise tests or unstructured generalized linear model (GLM) with regularization on vectorized networks to select edges distinguishing the groups, which ignore the network structure and make the results hard to interpret. In this paper, we develop a symmetric bilinear logistic regression (SBLR) with elastic-net penalty to identify a set of small clique subgraphs in network classification. Clique subgraphs, consisting of all the interconnections among a subset of brain regions, have appealing neurological interpretations as they may correspond to some anatomical circuits in the brain related to the outcome. We apply this method to study differences in the structural connectome between adolescents with high and low crystallized cognitive ability, using the crystallized cognition composite score, picture vocabulary and oral reading recognition tests from NIH Toolbox. A few clique subgraphs containing several small sets of brain regions are identified between different levels of functioning, indicating their importance in crystallized cognition.

neuroscience↗