arXiv subjects

Himadri S Samanta

Publications and source records attributed to Himadri S Samanta.

7 recordsLinked to original sources

arXiv2026

Recurrence-Based Nonlinear Vocal Dynamics as Digital Biomarkers for Depression Detection from Conversational Speech

Digital biomarkers for depression have largely relied on static acoustic descriptors, pooled summary statistics, or conventional machine learning representations. Such approaches may miss nonlinear temporal organization embedded in conversational vocal dynamics. We hypothesized that depression is associated with altered recurrence structure in vocal state trajectories, reflecting changes in how the vocal system revisits acoustic states over time. Using the depression subset of the DAIC-WOZ corpus with 142 labeled participants, we modeled frame-level COVAREP trajectories as nonlinear dynamical systems and derived recurrence-based biomarkers from 74 vocal channels. Logistic regression with feature selection and stratified cross-validation evaluated classification performance. Recurrence-based biomarkers achieved a mean cross-validated AUC of 0.689, exceeding static acoustic baselines, entropy-dynamics features, Hurst exponent features, determinism features, and Lyapunov-like instability proxies. Permutation testing indicated statistical significance with $p=0.004$. Pooled cross-validated predictions yielded AUC 0.665 with a 95\% bootstrap confidence interval of [0.568, 0.758]. These findings suggest that depression may be characterized by altered recurrence structure in conversational vocal dynamics and support nonlinear state-space analysis as a promising direction for digital psychiatric biomarkers.

cs.SD↗

arXiv2026

Entropy-Dominated Temporal Vocal Dynamics as Digital Biomarkers for Depression Detection

Automated depression detection often relies on static aggregation of conversational signals, potentially obscuring clinically meaningful behavioral dynamics. We investigated whether entropy-driven temporal biomarkers improve depression detection beyond standard pooled features using the DAIC-WOZ corpus. Using 142 labeled participants, we reconstructed utterance-level acoustic trajectories and compared pooled temporal baselines, trajectory dynamics, Shannon entropy biomarkers, recurrence quantification, sample entropy, fractal complexity, and coupling biomarkers under leakage-aware validation. Static pooling achieved an AUC of 0.593, trajectory dynamics improved performance to 0.637, and entropy biomarkers produced the strongest statistically significant improvement over pooled baselines (AUC 0.646; nested cross-validated AUC 0.615; permutation p = 0.017). Entropy biomarkers outperformed recurrence, coupling, sample entropy, and fractalbased features, with several biomarkers stable across folds. These findings suggest depression-related signal may lie less in average acoustic levels than in entropy of conversational dynamics, supporting temporally informed digital phenotypes for mental-health assessment.

q-bio.OT↗

arXiv2026

Grounded Multimodal Retrieval-Augmented Drafting of Radiology Impressions Using Case-Based Similarity Search

Automated radiology report generation has gained increasing attention with the rise of deep learning and large language models. However, fully generative approaches often suffer from hallucinations and lack clinical grounding, limiting their reliability in real-world workflows. In this study, we propose a multimodal retrieval-augmented generation (RAG) system for grounded drafting of chest radiograph impressions. The system combines contrastive image-text embeddings, case-based similarity retrieval, and citation-constrained draft generation to ensure factual alignment with historical radiology reports. A curated subset of the MIMIC-CXR dataset was used to construct a multimodal retrieval database. Image embeddings were generated using CLIP encoders, while textual embeddings were derived from structured impression sections. A fusion similarity framework was implemented using FAISS indexing for scalable nearest-neighbor retrieval. Retrieved cases were used to construct grounded prompts for draft impression generation, with safety mechanisms enforcing citation coverage and confidence-based refusal. Experimental results demonstrate that multimodal fusion significantly improves retrieval performance compared to image-only retrieval, achieving Recall@5 above 0.95 on clinically relevant findings. The grounded drafting pipeline produces interpretable outputs with explicit citation traceability, enabling improved trustworthiness compared to conventional generative approaches. This work highlights the potential of retrieval-augmented multimodal systems for reliable clinical decision support and radiology workflow augmentation

q-bio.QM↗

arXiv2026

Impact of automatic speech recognition quality on Alzheimer's disease detection from spontaneous speech: a reproducible benchmark study with lexical modeling and statistical validation

Early detection of Alzheimer's disease from spontaneous speech has emerged as a promising non-invasive screening approach. However, the influence of automatic speech recognition (ASR) quality on downstream clinical language modeling remains insufficiently understood. In this study, we investigate Alzheimer's disease detection using lexical features derived from Whisper ASR transcripts on the ADReSSo 2021 diagnosis dataset. We evaluate interpretable machine-learning models, including Logistic Regression and Linear Support Vector Machines, using TF-IDF text representations under repeated 5x5 stratified cross-validation. Our results demonstrate that transcript quality has a statistically significant impact on classification performance. Models trained on Whisper-small transcripts consistently outperform those using Whisper-base transcripts, achieving balanced accuracy above 0.7850 with Linear SVM. Paired statistical testing confirms that the observed improvements are significant. Importantly, classifier complexity contributes less to performance variation than ASR transcription quality. Feature analysis reveals that cognitively normal speakers produce more semantically precise object- and scene-descriptive language, whereas Alzheimer's speech is characterized by vagueness, discourse markers, and increased hesitation patterns. These findings suggest that high-quality ASR can enable simple, interpretable lexical models to achieve competitive Alzheimer's detection performance without explicit acoustic modeling. The study provides a reproducible benchmark pipeline and highlights ASR selection as a critical modeling decision in clinical speech-based artificial intelligence systems.

q-bio.QM↗

arXiv2021

Self-generated gradients guide cells in their long-distance expeditions

Self-generated gradients (SGG) provide robust steering cues that guide cells in their long-distance expeditions during embryonic development, immune response, and cancer metastasis. Cells generate their own local, dynamic gradients by breaking down the broadly distributed attractants in the environment, which leads to propulsion. How cells sense the environment and the interplay of self-generated force and long-range range chemotactic interactions determines the collective behavior of cells in vivo is largely unknown. We develop a theory to provide quantitative insight into how self-generated chemotaxis (SGC) steers the collective migration of cells during their long-distance journeys. In an externally imposed gradient, the cells exhibit super-diffusive motion at time less than persistent time ($τ_p$) and exhibit diffusive motion at time $t>τ_p$. On the other hand breakdown of chemo-attractants influences the density fluctuations and long-range correlations emerge in the density of chemotactic cells. The cells exhibit superdiffusive motion at a longtime limit. The SGG provides an effective mechanism for long-distance travel compared to an externally imposed gradient. Our theory provides a general framework for studying the non-equilibrium dynamics of collections of interacting particles.

cond-mat.soft↗

arXiv2019

Interstitial flows regulate collective cell migration heterogeneity through adhesion

The migration behaviors of cancer cells are known to be heterogeneous. However, the interplay between the adhesion interactions, dynamical shape changes and fluid flows in regulating cell migration heterogeneity and plasticity during cancer metastasis is still elusive. To further quantitative understanding of cell motility and morphology, we develop a theory using stochastic quantization method that describes the role of biophysical cues in regulating diverse cell motility. We show that the cumulative effect of time dependent adhesion interactions that determine the structural rearrangements and self-generated force due to actin remodeling, dictate the super-diffusive motion of mesenchymal phenotype in the absence of flow. Interstitial flows regulate cell motility phenotype and promote the amoeboid over mesenchymal motility through adhesion interactions. Cells exhibit a dynamical slowing down of collective migration, with a decreasing degree of super-diffusion. Mesenchymal cells are more persistent and diffusive compared to amoeboid cells. Our findings, suggest a mechanism of Interstitial flow induced directed motion of cancer cells through adhesion, and provide the much needed insight into a recent experimental observation concerning the diverse motility of breast cancer cells.

cond-mat.soft↗

arXiv2012

Exact solution of the Zwanzig-Lauritzen model of Polymer Crystallization under Tension

We solve a two dimensional model for polymer chain folding in the presence of mechanical pulling force ($f$) exactly using equilibrium statistical mechanics. Using analytically derived expression for the partition function we determine the phase diagram for the model in the $f$-temperature ($T$) plane. A square root singularity in the susceptibility indicates a second order phase transition from a folded to an unfolded state at a critical force ($f_c$) in the thermodynamic limit of infinitely long polymer chain. Surprisingly, the temperature dependence of $f_c$ shows a reentrant phase transition, which is reflected in an increase in $f_c$ as $T$ increases below a threshold value. For a range of $f$ values, the unfolded state is stable at both low and high temperatures. The high temperature unfolded state is stabilized by entropy whereas the low temperature unfolded state is dominated by favorable energy. The exact calculation could serve as a bench mark for testing approximate theories that are used in analyzing single molecule pulling experiments.

cond-mat.stat-mech↗