Brian Stasak

Dr. Brian Stasak

orcid

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

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

Representative Publications

View All Publications
11 March 2021 in Journal of Healthcare Informatics Research
Verified
Journal Article
Closed access
PDF file
BSBrian Stasak
Brian Stasak
ZHZhaocheng Huang
Zhaocheng Huang
SRSabah Razavi
Sabah Razavi
DJDale Joachim
Dale Joachim
JEJulien Epps
Julien Epps

1 September 2021 in Speech Communication
Verified
Journal Article
Closed access
BSBrian Stasak
Brian Stasak
JEJulien Epps
Julien Epps
HSHeather T. Schatten
Heather T. Schatten
IMIvan W. Miller
Ivan W. Miller
EPEmily Mower Provost
Emily Mower Provost
MAMichael F. Armey
Michael F. Armey

28 April 2022 in IEEE Pervasive Computing
Verified
Journal Article
Closed access
BSBrian Stasak
Brian Stasak
DJDale Joachim
Dale Joachim
JEJulien Epps
Julien Epps
Adults over the age of 60 years are a rising population at-risk for depression, and there is a need to create automatic screening for this illness. Most existing voice-based depression datasets comprise speakers younger than 60 and variations in speech due to age and depression are not well understood. In this study, which uses Patient Health Questionnaires for depression severity ground-truth, automatic depression detection is explored using acoustic-based prosodic, spectral, landmark, and voice quality features derived from smartphone recordings from 152 speakers in four different age ranges (e.g., 18–34, 35–48, 49–62, and 63–79). An age-dependent modeling paradigm for voice-based depression detection is proposed and evaluated. Results show that age-dependent models improve voice-based automatic depression classification accuracy with up to 10% absolute gains when compared with an age-agnostic model. Further, when compared with age-agnostic and gender-dependent models, age-dependent models often produced greater depressed class identification f-score sensitivity (up to 0.39 absolute gains). While automatically extracted acoustic voice features lead to statistically significant depression detection accuracy gains over the age-agnostic modeling baseline (4%–9% absolute), manually extracted voice quality features also are useful (4%–7% absolute gains over baseline). This investigation demonstrates the benefits of implementing age modeling to improve voice-based depression screening via smart devices.

1 November 2021 in Annual International Conference of the IEEE Engineering in Medicine and Biology Society
Verified
Conference Paper
Open access
BSBrian Stasak
Brian Stasak
ZHZhaocheng Huang
Zhaocheng Huang
JEJulien Epps
Julien Epps
DJDale Joachim
Dale Joachim
While the psychological Stroop color test has frequently been used to analyze response delays in temporal cognitive processing, minimal research has examined incorrect/correct verbal test response pattern differences exhibited in healthy control and clinically depressed populations. Further, the development of speech error features with an emphasis on sequential Stroop test responses has been unexplored for automatic depression classification. In this study which uses speech recorded via a smart device, an analysis of и-gram error sequence distributions shows that participants with clinical depression produce more Stroop color test errors, especially sequential errors, than the healthy controls. By utilizing и-gram error features derived from multisession manual transcripts, experimentation shows that trigram error features generate up to 95% depression classification accuracy, whereas an acoustic feature baseline achieve only upwards of 75%. Moreover, и-gram error features using ASR transcripts produced up to 90% depression classification accuracy.

9 August 2025 in Computer Speech & Language
Verified
Journal Article
Open access
BSBrian Stasak
Brian Stasak
JEJulien Epps
Julien Epps

21 August 2026 in Bioengineering
Verified
Journal Article
Open access
PDF file
BSBrian Stasak
Brian Stasak
JHJohn Holik
John Holik
DNDuy Duong Nguyen
Duy Duong Nguyen
TATomás Arias-Vergara
Tomás Arias-Vergara
Michael DöllingerMichael Döllinger
Michael Döllinger
CMCatherine Madill
Catherine Madill
Vocal rise time (VRT) acoustically measures from when vocal cord vibration begins to when phonation achieves a peak steady state. This acoustic-based study examines manual and automatic VRT measurements based on different vowel onset recordings (e.g., hard, modal, soft) from four datasets containing 146 adult Australian English-speaking females. Experiments test how the automatic Voice Onset Analysis Tool (VOAT) settings impact VRT measurement estimations. When compared to VOAT VRT baseline setting results, parameter adjustments using a Hilbert envelope 400 ms analysis segment with a low smoothing factor decreased VRT mean absolute error as low as 15.5 ms and 21.3 ms for modal and soft onsets, respectively. While the automated VOAT demonstrates objective repeatability with accelerated VRT extraction time advantages, still experiments herein show that the proposed adjusted settings are least accurate for hard onsets, whereby a mean absolute error as high as 71.0 ms is reported. Given three to five voice therapy sessions, females with voice disorders exhibited an 11% increase in VOAT VRT durations with reduced variation (6%), providing preliminary clinical evidence that this measure can assist in monitoring patients’ progress. This study reveals cross-corpora VRT norms per voice onset type, discusses VOAT parameter effects, and supports future automated VRT algorithm considerations.