An Emotion-Driven Vocal Biomarker-Based PTSD Screening Tool.

Goal: This paper introduces an automated post-traumatic stress disorder (PTSD) screening tool that could potentially be used as a self-assessment or inserted into routine medical visits to aid in PTSD diagnosis and treatment. Methods: With an emotion estimation algorithm providing arousal (excited to calm) and valence (pleasure to displeasure) levels through discourse, we select regions of the acoustic signal that are most salient for PTSD detection. Our algorithm was tested on a subset of data from the DVBIC-TBICoE TBI Study, which contains PTSD Check List Civilian (PCL-C) assessment scores. Results: Speech from low-arousal and positive-valence regions provide the highest discrimination for PTSD. Our model achieved an AUC (area under the curve) of 0.80 in detecting PCL-C ratings, outperforming models with no emotion filtering (AUC = 0.68). Conclusions: This result suggests that emotion drives the selection of the most salient temporal regions of an audio recording for PTSD detection.

The impact of recurrent laryngeal neuromonitoring on multi-dimensional voice outcomes following thyroid surgery.

BACKGROUND: Voice changes after thyroidectomy are common but not always related to recurrent laryngeal nerve (RLN) injury. We evaluated if RLN neuromonitoring correlated with non-RLN injury-related changes in voice after thyroidectomy. METHODS: Prospective multi-dimensional voice assessment was conducted on patients undergoing thyroidectomy before, 1-4 weeks, and 6 months postoperatively. Voice outcome (VO) was determined as normal (NormVO) or negative (NegVO) based upon combinations of patient-reported symptoms, videolaryngoscopy, a composite of acoustic measurements, and clinician-perceived voice quality. Groups with and without neuromonitoring were compared for early and durable differences in VO. RESULTS: Ninety-one patients underwent thyroidectomy; 39 with RLN neuromonitoring and 52 without. The two study groups were similar with regard to baseline characteristics including voice assessment. There was no difference in NegVO between neuromonitored and non-monitored patients at 1-4 weeks (n = 89; 32% vs. 27%; P = 0.81) and 6 months (n = 71, 14% vs. 7%; P = 0.42) after thyroidectomy. Neuromonitoring was associated with a 48-min increase in median operative time, but this finding was not statistically significant in a multivariate model. CONCLUSION: In this study, recurrent laryngeal neuromonitoring did not appear to influence non-RLN injury related VO as measured by a comprehensive multidimensional voice assessment.

A Rapid, Handheld Device to Assess Respiratory Resistance: Clinical and Normative Evidence.

INTRODUCTION: Following reports of respiratory symptoms among service members returning from deployment to South West Asia (SWA), an expert panel recommended pre-deployment spirometry be used to assess disease burden. Unfortunately, testing with spirometry is high cost and time-consuming. The airflow perturbation device (APD) is a handheld monitor that rapidly measures respiratory resistance (APD-Rr) and has promising but limited clinical data. Its speed and portability make it ideally suited for large volume pre-deployment screening. We conducted a pilot study to assess APD performance characteristics and develop normative values. MATERIALS AND METHODS: We prospectively enrolled subjects and derived reference equations for the APD from those without respiratory symptoms, pulmonary disease, or tobacco exposure. APD testing was conducted by medical technicians who received a 10-min in-service on its use. A subset of subjects performed spirometry and impulse oscillometry (iOS), administered by trained respiratory therapists. APD measures were compared with spirometry and iOS. RESULTS: The total study population included 199 subjects (55.8% males, body mass index 27.7 ± 6.0 kg/m2, age 49.9 ± 18.7 yr). Across the three APD trials, mean inspiratory (APD-Ri), expiratory (APD-Re), and average (APD-Ravg) resistances were 3.30 ± 1.0, 3.69 ± 1.2, and 3.50 ± 1.1 cm H2O/L/s. Reference equations were derived from 142 clinically normal volunteers. Height, weight, and body mass index were independently associated with APD-Ri, APD-Re, and APD-Ravg and were combined with age and gender in linear regression models. APD-Ri, APD-Re, and APD-Ravg were significantly inversely correlated with FEV1 (r = -0.39 to -0.42), FVC (r = -0.37 to -0.40), and FEF25-75 (r = -0.31 to -0.35) and positively correlated with R5 (r = 0.61-0.62), R20 (r = 0.50-0.52), X5 (r = -0.57 to -0.59), and FRES (r = 0.42-0.43). Bland-Altman plots showed that the APD-Rr closely approximates iOS when resistance is normal. CONCLUSION: Rapid testing was achieved with minimal training required, and reference equations were constructed. APD-Rr correlated moderately with iOS and weakly with spirometry. More testing is required to determine whether the APD has value for pre- and post-deployment respiratory assessment.

Does the presence or location of graphic markers affect untrained listeners' ratings of severity of dysphonia?

OBJECTIVE: To determine the effect of presence and location of severity labels for different types of visual analog scales (VAS) on overall severity (OS) ratings in dysphonic speech. STUDY DESIGN: Experimental, between group comparisons. METHODS: Dysphonic and normal voice samples from male and female speakers were presented to inexperienced listeners for judgments of OS. To rate samples, listeners used an undifferentiated 100-mm VAS labeled at the extremes, a VAS with nonlinearly distributed labels as in the "beta" version of the Consensus Auditory-Perceptual Evaluation of Voice (CAPE-V), or a VAS with symmetrically distributed labels as in the "official" version of the CAPE-V. RESULTS: Overall, mean OS ratings did not differ significantly across scale types, although ratings using the nonlinearly marked VAS were generally lower than those from other scales. This effect was significant for female speakers whose samples tended toward moderate OS. The ratings distribution, when compiled into 10-mm bins, differed significantly by scale type, with users of the nonlinearly marked scales skewing their ratings toward normal. CONCLUSIONS: The presence and placement of labels on VAS did not significantly affect OS ratings overall, but values were significantly lower when rating female voices with the nonlinearly labeled VAS. Results indicate that professionals should specify the scale type used for rating OS and use scales consistently when comparing voices.

Voice outcomes after total thyroidectomy, partial thyroidectomy, or non-neck surgery using a prospective multifactorial assessment.

BACKGROUND: Voice alteration remains a significant complication of thyroid surgery. We present a comparison of voice outcomes between total thyroidectomy (TT), partial thyroidectomy (PT), and non-neck (NN) surgery using a multifactorial voice-outcomes classification tool. STUDY DESIGN: Patients with normal voice (n = 112) were enrolled between July 2004 and March 2009. The patients underwent TT (n = 54), PT (n = 35), or NN (n = 23) surgery under general endotracheal anesthesia as part of a prospective observational study involving serial multimodality voice evaluation preoperatively, and at 2 weeks, 3 months, and 6 months postoperatively. Patients with adverse voice outcomes were grouped into the negative voice outcomes (NegVO) category, including patients with objective (abnormality on videolaryngostroboscopy and substantial voice dysfunction) and subjective (normal videolaryngostroboscopy but with notable voice impairment) NegVO. Voice outcomes were compared among study groups. RESULTS: Negative voice outcomes occurred in 46% (95% CI, 34-59%) and 14% (95% CI, 6-30%) of TT and PT groups, respectively. No NegVOs were observed after NN surgery. Early NegVOs were more common in the TT group than in the NN or PT groups (p < 0.001). Most voice disturbances resolved by 6 months (TT 84%; PT 92%) with no difference in NegVO among all groups (p = 0.23). Black race and significant changes in certain voice outcomes measures at the 2-week follow-up visit were identified as predictors of late (3 to 6 months) NegVO. CONCLUSIONS: This comprehensive voice outcomes study revealed that the extent of thyroidectomy impacts voice outcomes in the early postoperative period, and identified risk factors for late NegVO in post-thyroidectomy patients who should be considered for early voice rehabilitation referral.

Racial disparities in voice outcomes after thyroid and parathyroid surgery.

BACKGROUND: There is evidence that the outcomes of head and neck surgery may differ across racial and ethnic groups. Vocal changes related to the operation are an anticipated risk of thyroidectomy and parathyroidectomy. Race-specific voice outcomes after thyroid and parathyroid operations have not been reported. Therefore, our aim was to examine the potential disparity in voice outcomes between white and black patients after thyroid or parathyroid operations. PATIENTS AND METHODS: Eighty-seven patients (59 white and 28 black) were included in a prospective observational trial. Subjects were evaluated before operation, and 2 weeks, 3 months, and 6 months postoperatively using a comprehensive battery of functional voice assessments of voice characteristics. The association of race with voice outcomes over time was evaluated with generalized linear models. RESULTS: Aside from volume of pathologic specimen (black, 117.5 cm3 vs. white, 43.2 cm3; P = .004), presence of multinodular goiter (black, 32.1% vs. white, 6.8%; P = .004) or Hashimoto's thyroiditis (black, 3.6% vs. white, 28.8%; P = .009), there were no differences between racial groups. Blacks were more likely than whites to have negative voice outcomes (odds ratio, 2.6; 95% confidence interval, 1.1-6.2; P = .034] throughout the postoperative period, especially at 6 months (black, 25% vs. white, 4%; P = .018). This finding was related principally to divergent scores on the voice-related quality-of-life scale, the voice handicap index. CONCLUSION: We observed greater rates of self-reported, negative voice outcomes among blacks than whites after thyroid or parathyroid operations. The precise mechanism for this disparity has not been described. The observed racial disparity in self-perceived voice impairment in this study merits further investigation.

Prospective trial of voice outcomes after thyroidectomy: evaluation of patient-reported and clinician-determined voice assessments in identifying postthyroidectomy dysphonia.

BACKGROUND: Reliable voice grading systems to identify postoperative voice dysfunction by surgeons are needed. PURPOSE: To examine the utility of patient-reported and clinician-determined voice assessment in identifying postthyroidectomy voice dysfunction. PATIENTS AND METHODS: Fifty patients enrolled in a prospective observational trial evaluating voice function perioperatively by patient-reported symptoms (Voice Case History [VCHx]) and perceived voice handicap (Voice Handicap Index [VHI]), clinician-determined judgment of voice quality (Consensus Auditory-Perceptual Evaluation--Voice [CAPE-V]), and laryngeal examination via video laryngoscopy (VLS). Voice dysfunction at first postoperative visit in symptomatic patients was defined by objective laryngeal abnormalities on VLS. Postoperative changes from baseline in voice parameters were compared between patients with and without voice dysfunction using the Wilcoxon rank sum test. Receiver operating characteristics were evaluated to determine area under the curve (AUC) for tested parameters. RESULTS: Eight (16%) had early transient and 1 (2%) had permanent postoperative voice dysfunction. VCHx symptoms had negative (NPV) and positive (PPV) predictive values of 96%-100% and 39%-53%, respectively for voice dysfunction. The rating of overall severity from the CAPE-V was highly predictive (AUC = 0.96), and a change in severity from preoperative baseline >or=20% at 1-2 weeks had a PPV of 86% and NPV of 95% for postoperative dysphonia. Patient-reported total VHI score was most predictive (AUC = 0.97) and a change in VHI from preoperative baseline >or=25 early postoperatively had a PPV of 88% and NPV of 97% for postoperative dysphonia. CONCLUSION: Patient self-assessment of voice handicap using the VHI reliably identifies voice dysfunction after thyroidectomy. Patients with a change in VHI >or=25 from preoperative baseline warrant early referral to speech pathology and laryngology.