Sooyong Park

Mr. Sooyong Park

orcid

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

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

Representative Publications

View All Publications
20 June 2025 in npj Parkinson's Disease
Verified
Journal Article
Open access
PDF file
TATariq Adnan
Tariq Adnan
AAAbdelrahman Abdelkader
Abdelrahman Abdelkader
ZLZipei Liu
Zipei Liu
EHEkram Hossain
Ekram Hossain
SPSooyong Park
Sooyong Park
EHEhsan Hoque
Ehsan Hoque
We introduce a framework for screening Parkinson’s disease (PD) using English pangram utterances. Our dataset includes 1306 participants (392 with PD) from both home and clinical settings, covering diverse demographics (53.2% female). We used deep learning embeddings from Wav2Vec 2.0, WavLM, and ImageBind to capture speech dynamics indicative of PD. Our novel fusion model for PD classification aligns different speech embeddings into a cohesive feature space, outperforming baseline alternatives. In a stratified randomized split, the model achieved an AUROC of 88.9% and an accuracy of 85.7%. Statistical bias analysis showed equitable performance across sex, ethnicity, and age subgroups, with robustness across various disease durations and PD stages. Detailed error analysis revealed higher misclassification rates in specific age ranges for males and females, aligning with clinical insights. External testing yielded AUROCs of 82.1% and 78.4% on two clinical datasets, and an AUROC of 77.4% on an unseen general spontaneous English speech dataset, demonstrating versatility in natural speech analysis and potential for global accessibility and health equity.

19 December 2023 in Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies
Verified
Journal Article
Open access
WRWasifur Rahman
Wasifur Rahman
AAAbdelrahman Abdelkader
Abdelrahman Abdelkader
SLSangwu Lee
Sangwu Lee
PYPhillip Yang
Phillip Yang
SISaiful Islam
Saiful Islam
EHEhsan Hoque
Ehsan Hoque
We present a user-centric validation of a teleneurology platform, assessing its effectiveness in conveying screening information, facilitating user queries, and offering resources to enhance user empowerment. This validation process is implemented in the setting of Parkinson's disease (PD), in collaboration with a neurology department of a major medical center in the USA. Our intention is that with this platform, anyone globally with a webcam and microphone-equipped computer can carry out a series of speech, motor, and facial mimicry tasks. Our validation method demonstrates to users a mock PD risk assessment and provides access to relevant resources, including a chatbot driven by GPT, locations of local neurologists, and actionable and scientifically-backed PD prevention and management recommendations. We share findings from 91 participants (48 with PD, 43 without) aimed at evaluating the user experience and collecting feedback. Our framework was rated positively by 80.85% (standard deviation ± 8.92%) of the participants, and it achieved an above-average 70.42 (standard deviation ± 13.85) System-Usability-Scale (SUS) score. We also conducted a thematic analysis of open-ended feedback to further inform our future work. When given the option to ask any questions to the chatbot, participants typically asked for information about neurologists, screening results, and the community support group. We also provide a roadmap of how the knowledge generated in this paper can be generalized to screening frameworks for other diseases through designing appropriate recording environments, appropriate tasks, and tailored user-interfaces.

10 September 2023
Verified
Conference Paper
Open access
SISaiful Islam
Saiful Islam
SLSangwu Lee
Sangwu Lee
AAAbdelrahman Abdelkader
Abdelrahman Abdelkader
SPSooyong Park
Sooyong Park
EHEhsan Hoque
Ehsan Hoque
We present a web-based framework to screen for Parkinson’s disease (PD) by allowing users to perform neurological tests in their homes. Our web framework guides the users to complete three tasks involving speech, facial expression, and finger movements. The task videos are analyzed to classify whether the users show signs of PD. We present the results in an easy-to-understand manner, along with personalized resources to further access to treatment and care. Our framework is accessible by any major web browser, improving global access to neurological care.

18 August 2026 in Journal of Medical Internet Research
Verified
Journal Article
Open access
MIMd Saiful Islam
Md Saiful Islam
SPSooyong Park
Sooyong Park
EMEvelyn Xiaoxiao Ma
Evelyn Xiaoxiao Ma
TATariq Adnan
Tariq Adnan
EHEhsan Hoque
Ehsan Hoque
Background: Automated multimedia analysis of remotely recorded tasks offers a scalable approach to screening and remote monitoring of movement disorders such as Parkinson disease (PD). However, unsupervised recordings often suffer from quality issues that compromise model reliability. General multimedia quality checks may not detect task-specific failures, such as poor hand visibility during finger-tapping, inadequate facial framing during smile tasks, or background noise during speech tasks. Objective: This study aimed to develop and evaluate NeuroSift, a task-aware, interpretable machine learning framework for assessing recording quality and task compliance in home-recorded multimedia data collected for remote PD assessment. Methods: We analyzed 2516 home-recorded audio and video segments from 3 tasks: finger-tapping, facial expression (smile), and speech (pangram utterance). Three experts rated recordings as poor, borderline, or good quality. Task-specific annotation guidelines were developed through iterative review and discussion. We extracted interpretable features aligned with observable quality and compliance criteria and trained task-specific quality classification models. Interrater reliability was evaluated before and after guideline implementation using quadratic weighted Cohen κ (QWK), pairwise agreement, complete agreement, and intraclass correlation. Model performance was evaluated on a held-out test set (labeled using expert consensus) using accuracy, QWK, and ordinal classification accuracy (OCA, an accuracy metric that accounts for the ordered relationship among poor, borderline, and good labels). Feature importance was examined using Shapley additive explanations (SHAP), a method for estimating how individual features contribute to model predictions. Results: The task-specific guidelines significantly improved interrater reliability across all tasks (P<.001)—QWK increased from 0.46 to 0.89 for finger-tapping, from 0.61 to 0.84 for smile, and from 0.64 to 0.90 for speech. For 3-class quality classification, the best-performing models achieved QWK values of 0.71 for finger-tapping, 0.56 for smile, and 0.72 for speech. OCA was 82.1% for finger-tapping, 76.9% for smile, and 89.9% for speech. Most errors were adjacent-class errors, such as classifying poor recordings as borderline, whereas severe errors, such as classifying good recordings as poor, were rare. SHAP analyses identified task-specific sources of quality degradation that aligned with the annotation guidelines. Conclusions: NeuroSift was evaluated as a task-aware quality-classification framework for identifying low-quality or noncompliant multimedia data prior to downstream PD assessment. By combining structured quality guidelines, interpretable features, and explainable machine learning models, NeuroSift supports more transparent and user-correctable remote data collection. Although this study did not test downstream clinical impact, the framework offers a practical approach for quality-aware workflows that rely on user-recorded audio or video. This approach may extend beyond PD assessment to other remote settings, including telehealth, rehabilitation, and digital recruitment. Future studies should evaluate NeuroSift on independent datasets and examine its effects on downstream model performance, fairness, user experience, and workflow integration.

19 January 2026 in JMIR Preprints
Verified
Preprint
Open access
SISaiful Islam
Saiful Islam
SPSooyong Park
Sooyong Park
EMEvelyn Ma
Evelyn Ma
TATariq Adnan
Tariq Adnan
EHEhsan Hoque
Ehsan Hoque
BACKGROUND Movement disorders affect millions globally, yet access to specialists remains limited. Automated multimedia analysis offers a scalable path for screening and remote monitoring, but unsupervised recordings often suffer from quality issues that compromise model reliability. OBJECTIVE This study presents NeuroSift, a machine learning framework designed to automatically assess recording quality and task compliance in home-based multimedia data collected for movement disorder assessment. METHODS Three experts annotated 2,516 recordings corresponding to three common movement disorder tasks: finger tapping, facial expression, and speech. Based on these annotations, we developed task-specific quality guidelines, designed interpretable features, and trained quality classification models. Inter-rater agreement was evaluated before and after guideline implementation. Interpretable, task-specific features were extracted from each modality, and ordinal classification models were trained to categorize recordings as poor, borderline, or good quality. Model performance was evaluated against expert consensus, and feature importance analyses were conducted to identify common sources of quality degradation. RESULTS The quality guidelines significantly improved inter-rater agreement among experts (???? < 0.001). Across tasks, the quality classification models achieved 77–90% ordinal accuracy against expert consensus when classifying recordings as poor, borderline, or good quality. Feature importance analysis further identifies task-specific failures such as hand visibility, camera angle, or background noise. CONCLUSIONS NeuroSift is intended to strengthen trust in predictive models and support reliable deployment of multimedia-based neurological assessment tools at scale. By identifying low-quality or noncompliant recordings prior to downstream analysis, the framework would enhance the robustness, trustworthiness, and scalability of machine learning-based neurological assessment tools. CLINICALTRIAL None.

Verified
Journal Article
Open access
PDF file
MYMoon Youngbeen
Moon Youngbeen
HCHye Won Choi
Hye Won Choi
EKEun kyung Kang
Eun kyung Kang
JKJong K. Kim
Jong K. Kim
SPSooyong Park
Sooyong Park
KHKyeong-Man Hong
Kyeong-Man Hong
Genomic datasets generated using hg19 and hg38 are frequently integrated by coordinate liftover, but the correctability of associated discrepancies remains uncertain. We evaluated reciprocal liftover using experimentally validated allele pairs derived from homozygote hydatidiform mole DNA, heterozygote blood DNA, and seven physical dilution mixtures. Identical whole-exome sequencing FASTQ datasets were independently analyzed against hg19 and hg38 using DRAGEN v4.3, followed by reciprocal liftover and allele-pair comparison. Discordance occurred in 789/20,038 (3.94%) and 524/19,771 (2.65%) allele pairs for hg38→19 and hg19→38, respectively. Comprehensive reanalysis incorporating intronic variants identified 304 and 156 exon-to-intronic functional shifts, substantially reducing apparent null-null (N-N) shifts. Genotype flipping and functional-region shifts accounted for 722 (91.51%) and 404 (77.10%) of discordant allele pairs in the hg38→19 and hg19→38 comparisons, respectively, and were considered potentially correctable. Residual uncorrectable discrepancies comprised 67 and 120 allele pairs (0.33% and 0.61% of all informative allele pairs, respectively), including 35 and 83 N-N shifts. BAM-level inspection revealed heterogeneous discrepancies potentially involving reference-dependent alignment, local sequence representation, and variant-calling differences. These findings demonstrate that comprehensive variant harmonization can resolve many apparent cross-assembly discrepancies, although some remain unresolved. Reanalysis against a common reference assembly remains preferable when raw sequencing data are available.