IoT-Enabled WBAN and Machine Learning for Speech Emotion Recognition in Patients.

Internet of things (IoT)-enabled wireless body area network (WBAN) is an emerging technology that combines medical devices, wireless devices, and non-medical devices for healthcare management applications. Speech emotion recognition (SER) is an active research field in the healthcare domain and machine learning. It is a technique that can be used to automatically identify speakers' emotions from their speech. However, the SER system, especially in the healthcare domain, is confronted with a few challenges. For example, low prediction accuracy, high computational complexity, delay in real-time prediction, and how to identify appropriate features from speech. Motivated by these research gaps, we proposed an emotion-aware IoT-enabled WBAN system within the healthcare framework where data processing and long-range data transmissions are performed by an edge AI system for real-time prediction of patients' speech emotions as well as to capture the changes in emotions before and after treatment. Additionally, we investigated the effectiveness of different machine learning and deep learning algorithms in terms of performance classification, feature extraction methods, and normalization methods. We developed a hybrid deep learning model, i.e., convolutional neural network (CNN) and bidirectional long short-term memory (BiLSTM), and a regularized CNN model. We combined the models with different optimization strategies and regularization techniques to improve the prediction accuracy, reduce generalization error, and reduce the computational complexity of the neural networks in terms of their computational time, power, and space. Different experiments were performed to check the efficiency and effectiveness of the proposed machine learning and deep learning algorithms. The proposed models are compared with a related existing model for evaluation and validation using standard performance metrics such as prediction accuracy, precision, recall, F1 score, confusion matrix, and the differences between the actual and predicted values. The experimental results proved that one of the proposed models outperformed the existing model with an accuracy of about 98%.

A Machine Learning Approach to Screen for Otitis Media Using Digital Otoscope Images Labelled by an Expert Panel.

BACKGROUND: Otitis media includes several common inflammatory conditions of the middle ear that can have severe complications if left untreated. Correctly identifying otitis media can be difficult and a screening system supported by machine learning would be valuable for this prevalent disease. This study investigated the performance of a convolutional neural network in screening for otitis media using digital otoscopic images labelled by an expert panel. METHODS: Five experienced otologists diagnosed 347 tympanic membrane images captured with a digital otoscope. Images with a majority expert diagnosis (n = 273) were categorized into three screening groups Normal, Pathological and Wax, and the same images were used for training and testing of the convolutional neural network. Expert panel diagnoses were compared to the convolutional neural network classification. Different approaches to the convolutional neural network were tested to identify the best performing model. RESULTS: Overall accuracy of the convolutional neural network was above 0.9 in all except one approach. Sensitivity to finding ears with wax or pathology was above 93% in all cases and specificity was 100%. Adding more images to train the convolutional neural network had no positive impact on the results. Modifications such as normalization of datasets and image augmentation enhanced the performance in some instances. CONCLUSIONS: A machine learning approach could be used on digital otoscopic images to accurately screen for otitis media.

An Intra-Vehicular Wireless Multimedia Sensor Network for Smartphone-Based Low-Cost Advanced Driver-Assistance Systems.

Advanced driver-assistance system(s) (ADAS) are more prevalent in high-end vehicles than in low-end vehicles. Wired solutions of vision sensors in ADAS already exist, but are costly and do not cater for low-end vehicles. General ADAS use wired harnessing for communication; this approach eliminates the need for cable harnessing and, therefore, the practicality of a novel wireless ADAS solution was tested. A low-cost alternative is proposed that extends a smartphone's sensor perception, using a camera-based wireless sensor network. This paper presents the design of a low-cost ADAS alternative that uses an intra-vehicle wireless sensor network structured by a Wi-Fi Direct topology, using a smartphone as the processing platform. The proposed system makes ADAS features accessible to cheaper vehicles and investigates the possibility of using a wireless network to communicate ADAS information in a intra-vehicle environment. Other ADAS smartphone approaches make use of a smartphone's onboard sensors; however, this paper shows the application of essential ADAS features developed on the smartphone's ADAS application, carrying out both lane detection and collision detection on a vehicle by using wireless sensor data. A smartphone's processing power was harnessed and used as a generic object detector through a convolution neural network, using the sensory network's video streams. The network's performance was analysed to ensure that the network could carry out detection in real-time. A low-cost CMOS camera sensor network with a smartphone found an application, using Wi-Fi Direct, to create an intra-vehicle wireless network as a low-cost advanced driver-assistance system.

Low-Complexity Filter for Software-Defined Radio by Modulated Interpolated Coefficient Decimated Filter in a Hybrid Farrow.

Realising a low-complexity Farrow channelisation algorithm for multi-standard receivers in software-defined radio is a challenging task. A Farrow filter operates best at low frequencies while its performance degrades towards the Nyquist region. This makes wideband channelisation in software-defined radio a challenging task with high computational complexity. In this paper, a hybrid Farrow algorithm that combines a modulated Farrow filter with a frequency response interpolated coefficient decimated masking filter is proposed for the design of a novel filter with low computational complexity. A design example shows that the HFarrow filter bank achieved multiplier reduction of 50%, 70% and 64%, respectively, in comparison with non-uniform modulated discrete Fourier transform (NU MDFT FB), coefficient decimated filter bank (CD FB) and interpolated coefficient decimated (ICDM) filter algorithms. The HFarrow filter bank is able to provide the same number of sub-band channels as other algorithms such as non-uniform modulated discrete Fourier transform (NU MDFT FB), coefficient decimated filter bank (CD FB) and interpolated coefficient decimated (ICDM) filter algorithms, but with less computational complexity.

Diotic and Antiphasic Digits-in-noise Testing as a Hearing Screening and Triage Tool to Classify Type of Hearing Loss.

OBJECTIVES: The digits-in-noise test (DIN) is a popular self-test measure that has traditionally been used to screen for hearing loss by providing either a pass or refer result. Standard approaches either tested each ear monaurally or used a binaural diotic version where identical digits and noise were presented simultaneously to both ears. Recently, a dichotic, antiphasic version was developed, increasing sensitivity of the DIN to unilateral or asymmetric sensorineural hearing loss (SNHL) and conductive hearing loss (CHL). The purpose of this study was to determine predictors and normative ranges of the antiphasic and diotic DIN and to determine if a combination of diotic and antiphasic DIN could accurately categorize hearing into (1) normal, (2) bilateral SNHL, or (3) unilateral SNHL or CHL. DESIGN: The analytical sample consisted of 489 participants between the ages of 18 and 92 years with varying types, symmetry, and degrees of hearing loss. Degree and type of hearing loss were determined based on standard clinical four-frequency (0.5-4 kHz) pure-tone air and bone conduction threshold averages. The sample consisted of bilateral normal hearing (n = 293), bilateral SNHL (n = 172), unilateral SNHL (n = 42), and CHL (n = 32). All participants (n = 489) first completed an antiphasic DIN (digit stimuli 180° out-of-phase between ears), while 393 of the sample also completed a diotic DIN. Two procedures were assessed for their ability to categorize hearing into one of the three hearing groups. The first used a fixed antiphasic cutoff combined with a cutoff formed by a linear combination of antiphasic and diotic speech recognition threshold (SRT) or binaural intelligibility-level difference. RESULTS: Poorer ear pure-tone average was the strongest predictor of antiphasic DIN score, whereas better ear pure-tone average explained more of the variance in diotic SRT. The antiphasic DIN sensitivity and specificity was 90% and 84%, respectively, for detecting hearing loss, with outstanding area under the receiver operating characteristics values exceeding 0.93 to identify hearing loss in the poorer ear. The first fixed SRT cutoff procedure could categorize 75% of all participants correctly, while the second procedure increased correct categorization to 79%. False negative rates for both procedures were below 10%. CONCLUSIONS: A sequential antiphasic and diotic DIN could categorize hearing to a reasonable degree into three groups of (1) normal hearing; (2) bilateral SNHL; and (3) unilateral asymmetric SNHL or CHL. This type of approach could optimize care pathways using remote and contactless testing, by identifying unilateral SNHL and CHL as cases requiring medical referral. In contrast, bilateral SNHL cases could be referred directly to an audiologist, or nontraditional models like OTC hearing aids.

Listening Effort in School-Age Children With Normal Hearing Compared to Children With Limited Useable Hearing Unilaterally.

Objectives Children with limited hearing unilaterally might experience more listening effort than children with normal hearing, yet previous studies have not confirmed this. This study compared listening effort in school-age children with normal hearing and children with limited hearing unilaterally using behavioral and subjective listening effort measures. Design Two groups of school-age children (aged 7-12 years) participated: 19 with limited hearing unilaterally and 18 with normal hearing bilaterally. Participants completed digit triplet recognition tasks in quiet and in noise (-12 dB SNR) in three loudspeaker conditions: midline, direct, and indirect. Verbal response times during the recognition task were interpreted as behavioral listening effort. Subjective ratings of "task difficulty" and "hard to think" were interpreted as subjective listening effort. Participant age was included as a covariate in analysis of behavioral data. Results Noise negatively affected digit triplet recognition for both groups in the midline loudspeaker condition and for participants with limited hearing unilaterally in the direct and indirect conditions. Relative to their peers with normal hearing, children with limited hearing unilaterally exhibited significantly longer response times and higher ratings of effort only in the noisy, indirect condition. Differences between groups were evident even when age differences were controlled for statistically. Conclusions Using behavioral and subjective indices of listening effort, children with limited unilateral hearing demonstrated significantly more listening effort relative to their peers with normal hearing during the difficult indirect listening condition. Implications include classroom accommodations to limit indirect listening situations for children with limited useable hearing unilaterally and consideration of intervention options.

Listening Effort in School-Aged Children With Limited Useable Hearing Unilaterally: Examining the Effects of a Personal, Digital Remote Microphone System and a Contralateral Routing of Signal System.

Technology options for children with limited hearing unilaterally that improve the signal-to-noise ratio are expected to improve speech recognition and also reduce listening effort in challenging listening situations, although previous studies have not confirmed this. Employing behavioral and subjective indices of listening effort, this study aimed to evaluate the effects of two intervention options, remote microphone system (RMS) and contralateral routing of signal (CROS) system, in school-aged children with limited hearing unilaterally. Nineteen children (aged 7-12 years) with limited hearing unilaterally completed a digit triplet recognition task in three loudspeaker conditions: midline, monaural direct, and monaural indirect with three intervention options: unaided, RMS, and CROS system. Verbal response times were interpreted as a behavioral measure of listening effort. Participants provided subjective ratings immediately following behavioral measures. The RMS significantly improved digit triplet recognition across loudspeaker conditions and reduced verbal response times in the midline and indirect conditions. The CROS system improved speech recognition and listening effort only in the indirect condition. Subjective ratings analyses revealed that significantly more participants indicated that the remote microphone made it easier for them to listen and to stay motivated. Behavioral and subjective indices of listening effort indicated that an RMS provided the most consistent benefit for speech recognition and listening effort for children with limited unilateral hearing. RMSs could therefore be a beneficial technology option in classrooms for children with limited hearing unilaterally.

Pure-tone audiometry without bone-conduction thresholds: using the digits-in-noise test to detect conductive hearing loss.

Objective: COVID-19 has been prohibitive to traditional audiological services. No- or low-touch audiological assessment outside a sound-booth precludes test batteries including bone conduction audiometry. This study investigated whether conductive hearing loss (CHL) can be differentiated from sensorineural hearing loss (SNHL) using pure-tone air conduction audiometry and a digits-in-noise (DIN) test.Design: A retrospective sample was analysed using binomial logistic regressions, which determined the effects of pure tone thresholds or averages, speech recognition threshold (SRT), and age on the likelihood that participants had CHL or bilateral SNHL.Study sample: Data of 158 adults with bilateral SNHL (n = 122; PTA0.5-4 kHz > 25 dB HL bilaterally) or CHL (n = 36; air conduction PTA0.5-4 kHz > 25 dB HL and ≥20 dB air bone gap in the affected ears) were included.Results: The model which best discriminated between CHL and bilateral SNHL used low-frequency pure-tone average (PTA), diotic DIN SRT, and age with an area under the ROC curve of 0.98 and sensitivity and specificity of 97.2 and 93.4%, respectively.Conclusion: CHL can be accurately distinguished from SNHL using pure-tone air conduction audiometry and a diotic DIN. Restrictions on traditional audiological assessment due to COVID-19 require lower touch audiological care which reduces infection risk.

Listening Effort in Native and Nonnative English-Speaking Children Using Low Linguistic Single- and Dual-Task Paradigms.

Purpose It is not clear if behavioral indices of listening effort are sensitive to changes in signal-to-noise ratio (SNR) for young children (7-12 years old) from multilingual backgrounds. The purpose of this study was to explore the effects of SNR on listening effort in multilingual school-aged children (native English, nonnative English) as measured with a single- and a dual-task paradigm with low-linguistic speech stimuli (digits). The study also aimed to explore age effects on digit triplet recognition and response times (RTs). Method Sixty children with normal hearing participated, 30 per language group. Participants completed single and dual tasks in three SNRs (quiet, -10 dB, and -15 dB). Speech stimuli for both tasks were digit triplets. Verbal RTs were the listening effort measure during the single-task paradigm. A visual monitoring task was the secondary task during the dual-task paradigm. Results Significant effects of SNR on RTs were evident during both single- and dual-task paradigms. As expected, language background did not affect the pattern of RTs. The data also demonstrate a maturation effect for triplet recognition during both tasks and for RTs during the dual-task only. Conclusions Both single- and dual-task paradigms were sensitive to changes in SNR for school-aged children between 7 and 12 years of age. Language background (English as native language vs. English as nonnative language) had no significant effect on triplet recognition or RTs, demonstrating practical utility of low-linguistic stimuli for testing children from multilingual backgrounds.

Improving Sensitivity of the Digits-In-Noise Test Using Antiphasic Stimuli.

OBJECTIVES: The digits-in-noise test (DIN) has become increasingly popular as a consumer-based method to screen for hearing loss. Current versions of all DINs either test ears monaurally or present identical stimuli binaurally (i.e., diotic noise and speech, NoSo). Unfortunately, presentation of identical stimuli to each ear inhibits detection of unilateral sensorineural hearing loss (SNHL), and neither diotic nor monaural presentation sensitively detects conductive hearing loss (CHL). After an earlier finding of enhanced sensitivity in normally hearing listeners, this study tested the hypothesis that interaural antiphasic digit presentation (NoSπ) would improve sensitivity to hearing loss caused by unilateral or asymmetric SNHL, symmetric SNHL, or CHL. DESIGN: This cross-sectional study recruited adults (18 to 84 years) with various levels of hearing based on a 4-frequency pure-tone average (PTA) at 0.5, 1, 2, and 4 kHz. The study sample was comprised of listeners with normal hearing (n = 41; PTA ≤ 25 dB HL in both ears), symmetric SNHL (n = 57; PTA > 25 dB HL), unilateral or asymmetric SNHL (n = 24; PTA > 25 dB HL in the poorer ear), and CHL (n = 23; PTA > 25 dB HL and PTA air-bone gap ≥ 20 dB HL in the poorer ear). Antiphasic and diotic speech reception thresholds (SRTs) were compared using a repeated-measures design. RESULTS: Antiphasic DIN was significantly more sensitive to all three forms of hearing loss than the diotic DIN. SRT test-retest reliability was high for all tests (intraclass correlation coefficient r > 0.89). Area under the receiver operating characteristics curve for detection of hearing loss (>25 dB HL) was higher for antiphasic DIN (0.94) than for diotic DIN (0.77) presentation. After correcting for age, PTA of listeners with normal hearing or symmetric SNHL was more strongly correlated with antiphasic (rpartial[96] = 0.69) than diotic (rpartial = 0.54) SRTs. Slope of fitted regression lines predicting SRT from PTA was significantly steeper for antiphasic than diotic DIN. For listeners with normal hearing or CHL, antiphasic SRTs were more strongly correlated with PTA (rpartial[62] = 0.92) than diotic SRTs (rpartial[62] = 0.64). Slope of the regression line with PTA was also significantly steeper for antiphasic than diotic DIN. The severity of asymmetric hearing loss (poorer ear PTA) was unrelated to SRT. No effect of self-reported English competence on either antiphasic or diotic DIN among the mixed first-language participants was observed. CONCLUSIONS: Antiphasic digit presentation markedly improved the sensitivity of the DIN test to detect SNHL, either symmetric or asymmetric, while keeping test duration to a minimum by testing binaurally. In addition, the antiphasic DIN was able to detect CHL, a shortcoming of previous monaural or binaurally diotic DIN versions. The antiphasic DIN is thus a powerful tool for population-based screening. This enhanced functionality combined with smartphone delivery could make the antiphasic DIN suitable as a primary screen that is accessible to a large global audience.

Modernising speech audiometry: using a smartphone application to test word recognition.

OBJECTIVE: This study aimed to develop and assess a method to measure word recognition abilities using a smartphone application (App) connected to an audiometer. DESIGN: Word lists were recorded in South African English and Afrikaans. Analyses were conducted to determine the effect of hardware used for presentation (computer, compact-disc player, or smartphone) on the frequency content of recordings. An Android App was developed to enable presentation of recorded materials via a smartphone connected to the auxiliary input of the audiometer. Experiments were performed to test feasibility and validity of the developed App and recordings. STUDY SAMPLE: Participants were 100 young adults (18-30 years) with pure tone thresholds ≤15 dB across the frequency spectrum (250-8000 Hz). RESULTS: Hardware used for presentation had no significant effect on the frequency content of recordings. Listening experiments indicated good inter-list reliability for recordings in both languages, with no significant differences between scores on different lists at each of the tested intensities. Performance-intensity functions had slopes of 4.05%/dB for English and 4.75%/dB for Afrikaans lists at the 50% point. CONCLUSIONS: The developed smartphone App constitutes a feasible and valid method for measuring word recognition scores, and can support standardisation and accessibility of recorded speech audiometry.

The South African English Smartphone Digits-in-Noise Hearing Test: Effect of Age, Hearing Loss, and Speaking Competence.

OBJECTIVES: This study determined the effect of hearing loss and English-speaking competency on the South African English digits-in-noise hearing test to evaluate its suitability for use across native (N) and non-native (NN) speakers. DESIGN: A prospective cross-sectional cohort study of N and NN English adults with and without sensorineural hearing loss compared pure-tone air conduction thresholds to the speech reception threshold (SRT) recorded with the smartphone digits-in-noise hearing test. A rating scale was used for NN English listeners' self-reported competence in speaking English. This study consisted of 454 adult listeners (164 male, 290 female; range 16 to 90 years), of whom 337 listeners had a best ear four-frequency pure-tone average (4FPTA; 0.5, 1, 2, and 4 kHz) of ≤25 dB HL. RESULTS: A linear regression model identified three predictors of the digits-in-noise SRT, namely, 4FPTA, age, and self-reported English-speaking competence. The NN group with poor self-reported English-speaking competence (≤5/10) performed significantly (p < 0.01) poorer than the N and NN (≥6/10) groups on the digits-in-noise test. Screening characteristics of the test improved with separate cutoff values depending on English-speaking competence for the N and NN groups (≥6/10) and NN group alone (≤5/10). Logistic regression models, which include age in the analysis, showed a further improvement in sensitivity and specificity for both groups (area under the receiver operating characteristic curve, 0.962 and 0.903, respectively). CONCLUSIONS: Self-reported English-speaking competence had a significant influence on the SRT obtained with the smartphone digits- in-noise test. A logistic regression approach considering SRT, self-reported English-speaking competence, and age as predictors of best ear 4FPTA >25 dB HL showed that the test can be used as an accurate hearing screening tool for N and NN English speakers. The smartphone digits-in-noise test, therefore, allows testing in a multilingual population familiar with English digits using dynamic cutoff values that can be chosen according to self-reported English-speaking competence and age.

Automated Smartphone Threshold Audiometry: Validity and Time Efficiency.

BACKGROUND: Smartphone-based threshold audiometry with automated testing has the potential to provide affordable access to audiometry in underserved contexts. PURPOSE: To validate the threshold version (hearTest) of the validated hearScreen™ smartphone-based application using inexpensive smartphones (Android operating system) and calibrated supra-aural headphones. RESEARCH DESIGN: A repeated measures within-participant study design was employed to compare air-conduction thresholds (0.5-8 kHz) obtained through automated smartphone audiometry to thresholds obtained through conventional audiometry. STUDY SAMPLE: A total of 95 participants were included in the study. Of these, 30 were adults, who had known bilateral hearing losses of varying degrees (mean age = 59 yr, standard deviation [SD] = 21.8; 56.7% female), and 65 were adolescents (mean age = 16.5 yr, SD = 1.2; 70.8% female), of which 61 had normal hearing and the remaining 4 had mild hearing losses. DATA ANALYSIS: Threshold comparisons were made between the two test procedures. The Wilcoxon signed-ranked test was used for comparison of threshold correspondence between manual and smartphone thresholds and the paired samples t test was used to compare test time. RESULTS: Within the adult sample, 94.4% of thresholds obtained through smartphone and conventional audiometry corresponded within 10 dB or less. There was no significant difference between smartphone (6.75-min average, SD = 1.5) and conventional audiometry test duration (6.65-min average, SD = 2.5). Within the adolescent sample, 84.7% of thresholds obtained at 0.5, 2, and 4 kHz with hearTest and conventional audiometry corresponded within ≤5 dB. At 1 kHz, 79.3% of the thresholds differed by ≤10 dB. There was a significant difference (p < 0.01) between smartphone (7.09 min, SD = 1.2) and conventional audiometry test duration (3.23 min, SD = 0.6). CONCLUSIONS: The hearTest application with calibrated supra-aural headphones provides a cost-effective option to determine valid air-conduction hearing thresholds.

Smartphone-Based Hearing Screening at Primary Health Care Clinics.

OBJECTIVE: To evaluate the performance of smartphone-based hearing screening with the hearScreen application in terms of sensitivity, specificity, referral rates, and time efficiency at two primary health care clinics. DESIGN: Nonprobability purposive sampling was used at both clinics. A total of 1236 participants (mean age: 37.8 ± SD 17.9 and range 3 to 97 years; 71.3% female) were included in the final analysis. Participants were screened using the hearScreen application following a two-step screening protocol and diagnostic pure-tone audiometry to confirm hearing status. RESULTS: Sensitivity and specificity for smartphone screening was 81.7 and 83.1%, respectively, with a positive and negative predictive value of 87.6 and 75.6%, respectively. Sex [χ(1, N = 126) = 0.304, p > 0.05] and race [χ(1, N = 126) = 0.169, p > 0.05)] had no significant effect on screening outcome for children while for adults age (p < 0.01; ß = 0.04) and sex (p = 0.02; ß = -0.53) had a significant effect on screening outcomes with males more likely to fail. Overall referral rate across clinics was 17.5%. Initial screening test times were less than a minute (48.8 seconds ± 20.8 SD) for adults and slightly more than a minute for children (73.9 seconds ± 44.5 SD). CONCLUSIONS: The hearScreen smartphone application provides time-efficient identification of hearing loss with adequate sensitivity and specificity for accurate testing at primary health care settings.

Otitis Media Diagnosis for Developing Countries Using Tympanic Membrane Image-Analysis.

BACKGROUND: Otitis media is one of the most common childhood diseases worldwide, but because of lack of doctors and health personnel in developing countries it is often misdiagnosed or not diagnosed at all. This may lead to serious, and life-threatening complications. There is, thus a need for an automated computer based image-analyzing system that could assist in making accurate otitis media diagnoses anywhere. METHODS: A method for automated diagnosis of otitis media is proposed. The method uses image-processing techniques to classify otitis media. The system is trained using high quality pre-assessed images of tympanic membranes, captured by digital video-otoscopes, and classifies undiagnosed images into five otitis media categories based on predefined signs. Several verification tests analyzed the classification capability of the method. FINDINGS: An accuracy of 80.6% was achieved for images taken with commercial video-otoscopes, while an accuracy of 78.7% was achieved for images captured on-site with a low cost custom-made video-otoscope. INTERPRETATION: The high accuracy of the proposed otitis media classification system compares well with the classification accuracy of general practitioners and pediatricians (~64% to 80%) using traditional otoscopes, and therefore holds promise for the future in making automated diagnosis of otitis media in medically underserved populations.

Smartphone threshold audiometry in underserved primary health-care contexts.

OBJECTIVE: To validate a calibrated smartphone-based hearing test in a sound booth environment and in primary health-care clinics. DESIGN: A repeated-measure within-subject study design was employed whereby air-conduction hearing thresholds determined by smartphone-based audiometry was compared to conventional audiometry in a sound booth and a primary health-care clinic environment. STUDY SAMPLE: A total of 94 subjects (mean age 41 years ± 17.6 SD and range 18-88; 64% female) were assessed of whom 64 were tested in the sound booth and 30 within primary health-care clinics without a booth. RESULTS: In the sound booth 63.4% of conventional and smartphone thresholds indicated normal hearing (≤15 dBHL). Conventional thresholds exceeding 15 dB HL corresponded to smartphone thresholds within ≤10 dB in 80.6% of cases with an average threshold difference of -1.6 dB ± 9.9 SD. In primary health-care clinics 13.7% of conventional and smartphone thresholds indicated normal hearing (≤15 dBHL). Conventional thresholds exceeding 15 dBHL corresponded to smartphone thresholds within ≤10 dB in 92.9% of cases with an average threshold difference of -1.0 dB ± 7.1 SD. CONCLUSIONS: Accurate air-conduction audiometry can be conducted in a sound booth and without a sound booth in an underserved community health-care clinic using a smartphone.

Smartphone hearing screening in mHealth assisted community-based primary care.

INTRODUCTION: Access to ear and hearing health is a challenge in developing countries, where the burden of disabling hearing loss is greatest. This study investigated community-based identification of hearing loss using smartphone hearing screening (hearScreen™) operated by community health workers (CHWs) in terms of clinical efficacy and the reported experiences of CHWs. METHOD: The study comprised two phases. During phase one, 24 CHWs performed community-based hearing screening as part of their regular home visits over 12 weeks in an underserved community, using automated test protocols employed by the hearScreen™ smartphone application, operating on low-cost smartphones with calibrated headphones. During phase two, CHWs completed a questionnaire regarding their perceptions and experiences of the community-based screening programme. RESULTS: Data analysis was conducted on the results of 108 children (2-15 years) and 598 adults (16-85 years). Referral rates for children and adults were 12% and 6.5% respectively. Noise exceeding permissible levels had a significant effect on screen results at 25 dB at 1 kHz (p<0.05). Age significantly affected adult referral rates (p < 0.05), demonstrating a lower rate (4.3%) in younger as opposed to older adults (13.2%). CHWs were positive regarding the hearScreen™ solution in terms of usability, need for services, value to community members and time efficiency. CONCLUSION: Smartphone-based hearing screening allows CHWs to bring hearing health care to underserved communities at a primary care level. Active noise monitoring and data management features allow for quality control and remote monitoring for surveillance and follow-up.

Clinical Validity of hearScreen™ Smartphone Hearing Screening for School Children.

OBJECTIVES: The study aimed to determine the validity of a smartphone hearing screening technology (hearScreen™) compared with conventional screening audiometry in terms of (1) sensitivity and specificity, (2) referral rate, and (3) test time. DESIGN: One thousand and seventy school-age children in grades 1 to 3 (8 ± 1.1 average years) were recruited from five public schools. Children were screened twice, once using conventional audiometry and once with the smartphone hearing screening. Screening was conducted in a counterbalanced sequence, alternating initial screen between conventional or smartphone hearing screening. RESULTS: No statistically significant difference in performance between techniques was noted, with smartphone screening demonstrating equivalent sensitivity (75.0%) and specificity (98.5%) to conventional screening audiometry. While referral rates were lower with the smartphone screening (3.2 vs. 4.6%), it was not significantly different (p > 0.05). Smartphone screening (hearScreen™) was 12.3% faster than conventional screening. CONCLUSION: Smartphone hearing screening using the hearScreen™ application is accurate and time efficient.

A Review of Intelligent Driving Style Analysis Systems and Related Artificial Intelligence Algorithms.

In this paper the various driving style analysis solutions are investigated. An in-depth investigation is performed to identify the relevant machine learning and artificial intelligence algorithms utilised in current driver behaviour and driving style analysis systems. This review therefore serves as a trove of information, and will inform the specialist and the student regarding the current state of the art in driver style analysis systems, the application of these systems and the underlying artificial intelligence algorithms applied to these applications. The aim of the investigation is to evaluate the possibilities for unique driver identification utilizing the approaches identified in other driver behaviour studies. It was found that Fuzzy Logic inference systems, Hidden Markov Models and Support Vector Machines consist of promising capabilities to address unique driver identification algorithms if model complexity can be reduced.

Development and validation of a smartphone-based digits-in-noise hearing test in South African English.

OBJECTIVE: The objective of this study was to develop and validate a smartphone-based digits-in-noise hearing test for South African English. DESIGN: Single digits (0-9) were recorded and spoken by a first language English female speaker. Level corrections were applied to create a set of homogeneous digits with steep speech recognition functions. A smartphone application was created to utilize 120 digit-triplets in noise as test material. An adaptive test procedure determined the speech reception threshold (SRT). Experiments were performed to determine headphones effects on the SRT and to establish normative data. STUDY SAMPLE: Participants consisted of 40 normal-hearing subjects with thresholds ≤15 dB across the frequency spectrum (250-8000 Hz) and 186 subjects with normal-hearing in both ears, or normal-hearing in the better ear. RESULTS: The results show steep speech recognition functions with a slope of 20%/dB for digit-triplets presented in noise using the smartphone application. The results of five headphone types indicate that the smartphone-based hearing test is reliable and can be conducted using standard Android smartphone headphones or clinical headphones. CONCLUSION: A digits-in-noise hearing test was developed and validated for South Africa. The mean SRT and speech recognition functions correspond to previous developed telephone-based digits-in-noise tests.