Vocal Acoustic Biomarkers in Depression

This study will incorporate IVR data collection procedures developed in a prior pilot study
into a multi-site industry-sponsored Phase 4 randomized, double-blind, placebo-controlled
study (NCT00406952) using an established antidepressant at a therapeutically efficacious
dose. The analytic procedures used to extract the vocal acoustic measures will be automated
to facilitate near rapid turn around between speech sample acquisition and delivery of
biomarker outcomes data. Multivariate models of vocal acoustic information to enhance the
sensitivity of treatment response measurement and prediction of patient response likelihood
will be further refined.

A total of 150 subjects will be recruited in the proposed studies, a sample of 50 from each
of three targeted depressed patient populations - young adults (18-25 years old), elderly
patients (60 years and older), and recent Asian immigrants with financial, linguistic, and
cultural barriers to health care access. Eligible subjects will meet diagnostic criteria
for Major Depressive Disorder, provide written informed consent, and will have baseline
depression severity scores of 10 or greater on the QIDS. Pregnant women, women of child
bearing potential not using a medically accepted means of contraception, suicidal or
homicidal patients, or those with unstable medical illness including cardiovascular,
hepatic, renal, respiratory, endocrine, neurological, hematological disease, or with
clinical or laboratory evidence of hypothyroidism will be excluded. Patients beginning a
new treatment for depression will be clinically evaluated (QIDS-C) during face-to-face
interviews at the beginning of treatment (Baseline), and followed up 4 (study midpoint) and
8 weeks later (end-point). After completing the clinical interviews study participants will
complete a series of IVR-based assessments using a standard desktop touch-tone telephone.