Thasina Tabashum

Mrs. Thasina Tabashum


Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
17 May 2024 in JMIR Medical Informatics
Verified
Journal Article
Open access
TTThasina Tabashum
Thasina Tabashum
RSRobert Cooper Snyder
Robert Cooper Snyder
MOMegan O'Brien
Megan O'Brien
MAMark V. Albert
Mark V. Albert
Background: With the increasing availability of data, computing resources, and easier-to-use software libraries, machine learning (ML) is increasingly used in disease detection and prediction, including for Parkinson disease (PD). Despite the large number of studies published every year, very few ML systems have been adopted for real-world use. In particular, a lack of external validity may result in poor performance of these systems in clinical practice. Additional methodological issues in ML design and reporting can also hinder clinical adoption, even for applications that would benefit from such data-driven systems. Objective: To sample the current ML practices in PD applications, we conducted a systematic review of studies published in 2020 and 2021 that used ML models to diagnose PD or track PD progression. Methods: We conducted a systematic literature review in accordance with PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines in PubMed between January 2020 and April 2021, using the following exact string: “Parkinson’s” AND (“ML” OR “prediction” OR “classification” OR “detection” or “artificial intelligence” OR “AI”). The search resulted in 1085 publications. After a search query and review, we found 113 publications that used ML for the classification or regression-based prediction of PD or PD-related symptoms. Results: Only 65.5% (74/113) of studies used a holdout test set to avoid potentially inflated accuracies, and approximately half (25/46, 54%) of the studies without a holdout test set did not state this as a potential concern. Surprisingly, 38.9% (44/113) of studies did not report on how or if models were tuned, and an additional 27.4% (31/113) used ad hoc model tuning, which is generally frowned upon in ML model optimization. Only 15% (17/113) of studies performed direct comparisons of results with other models, severely limiting the interpretation of results. Conclusions: This review highlights the notable limitations of current ML systems and techniques that may contribute to a gap between reported performance in research and the real-life applicability of ML models aiming to detect and predict diseases such as PD.

10 October 2020 in Journal of Healthcare Engineering
Verified
Journal Article
Open access
PDF file
SZSamuel Zelman
Samuel Zelman
MDMichael Dow
Michael Dow
TTThasina Tabashum
Thasina Tabashum
Ting XiaoTing Xiao
Ting Xiao
MAMark V. Albert
Mark V. Albert
Measuring physical activity using wearable sensors is essential for quantifying adherence to exercise regiments in clinical research and motivating individuals to continue exercising. An important aspect of wearable activity tracking is counting particular movements. One limitation of many previous models is the need to design the counting for a specific exercise. However, during physical therapy, some movements are unique to the patient and also valuable to track. To address this, we create an automatic repetition counting system that is flexible enough to measure multiple distinct and repeating movements during physical therapy without being trained on the specific motion. Accelerometers, using smartphones, were attached to the body or held by participants to track repetitive motions during different exercises. 18 participants completed a series of 10 exercises for 30 seconds, including arm circles, bicep curls, bridges, sit-ups, elbow extensions, leg lifts, lunges, push-ups, squats, and upper trunk rotations. To count the repetitions of each exercise, we apply three analysis techniques: (a) threshold crossing, (b) threshold crossing with a low-pass filter, and (c) Fourier transform. The results demonstrate that arm circles and push-ups can be tracked well, while less periodic and irregular motions such as upper trunk rotations are more difficult. Overall, threshold crossing with low-pass filtering achieves the best performance among these methods. We conclude that the proposed automatic counting system is capable of tracking exercise repetition without prior training and development for that activity.

1 July 2019
Verified
Conference Paper
Closed access
TTThasina Tabashum
Thasina Tabashum
ASAbdul Mutalab Shaykat
Abdul Mutalab Shaykat
SASheikh Abujar
Sheikh Abujar
MOMohibullah
Mohibullah
SCShekhor Chanda
Shekhor Chanda
Emotion tagging aims to tag words with appropriate emotion that a word expresses in a document or a sentence. In this paper, authors introduced an approach to tag words automatically and then analyzed in sentence and document level. The evaluation of tagged words described and the performance based on changing the tone of a document has been explained. The experiment showed that the approach could be effective if the tag words are not constant or static. Application of this approach in a contextual level analysis is briefly concluded and proposed tentative emotional tree by using emotion tagging in word level.

18 October 2022 in Bioengineering
Verified
Journal Article
Open access
PDF file
TTThasina Tabashum
Thasina Tabashum
Ting XiaoTing Xiao
Ting Xiao
CJChandrasekaran Jayaraman
Chandrasekaran Jayaraman
CMChaithanya K. Mummidisetty
Chaithanya K. Mummidisetty
AJArun Jayaraman
Arun Jayaraman
MAMark V. Albert
Mark V. Albert
We created an overall assessment metric using a deep learning autoencoder to directly compare clinical outcomes in a comparison of lower limb amputees using two different prosthetic devices—a mechanical knee and a microprocessor-controlled knee. Eight clinical outcomes were distilled into a single metric using a seven-layer deep autoencoder, with the developed metric compared to similar results from principal component analysis (PCA). The proposed methods were used on data collected from ten participants with a dysvascular transfemoral amputation recruited for a prosthetics research study. This single summary metric permitted a cross-validated reconstruction of all eight scores, accounting for 83.29% of the variance. The derived score is also linked to the overall functional ability in this limited trial population, as improvements in each base clinical score led to increases in this developed metric. There was a highly significant increase in this autoencoder-based metric when the subjects used the microprocessor-controlled knee (p < 0.001, repeated measures ANOVA). A traditional PCA metric led to a similar interpretation but captured only 67.3% of the variance. The autoencoder composite score represents a single-valued, succinct summary that can be useful for the holistic assessment of highly variable, individual scores in limited clinical datasets.

14 December 2025 in Biocomputing 2020
Verified
Conference Paper
Open access
JRJubair Ibn Malik Rifat
Jubair Ibn Malik Rifat
TTThasina Tabashum
Thasina Tabashum
MRMarufi Rahman
Marufi Rahman
FMFarhad Mokter
Farhad Mokter
SESarthak Engala
Sarthak Engala
SBSerdar Bozdag
Serdar Bozdag
Biologically informed deep neural networks, which connect input layer to hidden layers based on genepathway relationship have gained popularity in recent years. However, most existing methods do not incorporate protein-protein interactions (PPI) and protein-DNA interactions (PDI) in their designs. In this study, we introduce BioLM-NET, a deep learning-based framework that fuses single cell or bulk gene expression data and DNA methylation data with prior biological knowledge including Protein- Protein Interactions (PPI), Protein-DNA Interactions (PDI). BioLM-NET also aggregates latent representation of omics signals at pathway-level through an attention-based pathway layer where a pretrained large language model (LLM) was incorporated to generate context-specific gene embeddings. We evaluated BioLM-NET on single cell colorectal cancer data from scTrioseq2 platform to predict primary and metastatic cancer cells, on TCGA-BRCA, TCGA-GBM, TCGA-COAD to predict cancer subtypes and ROSMAP data to predict Alzheimer's disease patient. Our results showed that BioLMNET outperformed baseline and state-of-the-art (SOTA) methods, P-NET and PASNet with statistical significance on scTrioseq2 data, TCGA-COAD and ROSMAP data and ties with SVM and Dense neural network on TCGA-BRCA data. Our ablation studies demonstrated the importance of incorporating PPI, PDI data and attention-based pathway layer. We also interpret our models and found out that our important input features are significantly enriched in GO terms and KEGG pathways and can serve as potential biomarkers or therapeutic targets for the corresponding disease.

1 June 2026 in JMIR Preprints
Verified
Preprint
Open access
FRFahmida Yasmin Rifat
Fahmida Yasmin Rifat
TTThasina Tabashum
Thasina Tabashum
JKJoseph J Krzak
Joseph J Krzak
KKKaren Kruger
Karen Kruger
AGAdam Graf
Adam Graf
MAMark V. Albert
Mark V. Albert
BACKGROUND Pain is common and clinically consequential in children with cerebral palsy (CP), but retrospective pain severity assessment from electronic health records (EHRs) is difficult. Pain evidence is often dispersed across long narrative notes, many children have limited ability to self-report, and structured pain scores may be influenced by proxy reporting and encounter context. OBJECTIVE This study introduces the first systematic benchmark for three-class pain severity classification from clinical notes of children with CP and evaluates whether modern Natural Language Processing (NLP) and large language model approaches can recover clinically meaningful severity signals from real-world EHR documentation. METHODS We conducted a retrospective study using 15,969 de-identified clinical notes from 1,467 children with CP treated across the Shriners Children’s Hospital Network over 19 years. Notes were linked to encounter-level numeric pain ratings and grouped into No/Mild (0–3), Moderate (4–6), and Severe (7–10) classes. Using patient-stratified splits, we compared supervised encoder-fusion models, explanation-augmented local language models, frozen instruction-following LLMs, and a late-fusion ensemble. RESULTS Performance improved as models incorporated broader context and complementary decision patterns. The best supervised encoder-fusion model achieved a macro F1 of 0.600, while explanation-augmented local LLMs improved to 0.620 when clinical notes were combined with generated explanations. Few-shot GPT-5 was the strongest standalone model in this held-out test set, achieving a macro F1 of 0.644. The best overall performance was achieved by a late-fusion ensemble of Bio-Mistral-7B, zero-shot GPT-5, and few-shot GPT-5, reaching a macro F1 of 0.66. SHapley Additive exPlanations (SHAP) analysis showed that No/Mild predictions were driven mainly by negated pain language, whereas Moderate and Severe predictions were associated with explicit pain terms and severity cues. CONCLUSIONS Modern NLP and LLM-based approaches can extract clinically meaningful pain severity signals from real-world EHR notes of children with CP. Performance was strongest when models used long-context modeling, explanation-guided supervision, and complementary prediction patterns. Remaining errors reflected the subjective, proxy-mediated, and temporally variable nature of pain documentation. Therefore, these models should support retrospective pain research, cohort identification, and documentation review. To our knowledge, this is the first study to benchmark multi-level pain severity classification from unstructured EHR notes of children with CP.