2019 NHCA Annual Conference

The world is increasingly sensed and measured by networked systems, opening up incredible opportunities for expanding human knowledge. While sensory perception remains bounded to our immediate surroundings, AI systems take in vast quantities of data and provide insight on a global scale. Naturally then, we are spending more time consuming, sharing, and acting on high-level machine analyses. However, if we grow to rely exclusively on discrete knowledge distilled from data, we are in danger of dulling our own perceptual sensitivities. In this talk, I lay out a future in which the sensor networks that blanket our world become extensions of human perception, forming the basis for a restoration of sensitivity as well as a transformative augmentation of memory and cognition. To ground this vision, I introduce a wetland restoration project called the Tidmarsh Living Observatory, which has played host to a series of novel environmental monitoring systems, sensory wearables, and AI interfaces. Using my and others’ work as examples, I argue that non-visual perceptual modalities, and hearing specifically, will underpin an imminent revolution in human-computer interaction, and vastly enrich the positive transformative potential of AI. Our perceptual sensitivities provide us the means to acquire new knowledge. They guide our attention, spark curiosity and stir emotion, fostering the kind of intuitions, sensibilities, and care on which action for the future is predicated. What does climate change sound like?

Industrial hygienists have long recognized that excessive occupational noise exposure is a health risk for noise-induced hearing loss. However, in the past several decades, noise exposures both inside and outside the workplace have been potentially linked to additional health outcomes. With the objective of identifying safe levels of exposure, the Centers for Disease Control and Prevention (CDC) National Center for Environmental Health has conducted a formal review process to better understand the range of health risks potentially resulting from noise. To accomplish this objective, eleven systematic reviews have been conducted to identify what noise levels, and for how long, are associated with eleven specific health end points: 1) hearing loss; 2) ischemic heart disease; 3) hypertension; 4) mental health/psychological disorders; 5) injuries; 6) sleep disturbance; 7) low birth weight/premature birth; 8) endocrine disruption; 9) cognitive impairment; 10) obesity/overweight, and; 11) cancer or tumorigenesis. Each review involved a systematic literature review and rating of the available evidence; reviews also included a meta-analysis of health risks when sufficient evidence was available. Using these reviews, CDC will be able to raise awareness and enhance prevention strategies.

The accurate classification of nonverbal human produced audio events opens the door to numerous applications beyond health monitoring. Voluntary events, such as tongue clicking and teeth chattering, may lead to a novel way of silent interface command. Involuntary events, such as coughing and clearing the throat, may advance the current state-of-the-art in hearing health research. The challenge of such applications is the balance between the processing capabilities of a small intra-aural device and the accuracy of classification. In this pilot study, 10 nonverbal audio events are captured inside the ear canal blocked by an intra-aural device.  The performance of three classifiers is investigated:  Gaussian Mixture Model (GMM), Support Vector Machine and  Multi-Layer Perceptron. Each classifier is trained using three different feature vector structures constructed using the mel-frequency cepstral (MFCC) coefficients and their derivatives. Fusion of the MFCCs with the auditory-inspired amplitude modulation features (AAMF) is also investigated. Classification is compared between binaural and monaural training sets as well as for noisy and clean conditions. The highest accuracy is achieved at 75.45% using the GMM classifier with the binaural MFCC+AAMF clean training set. Accuracy of 73.47% is achieved by training and testing the classifier with the binaural clean and noisy dataset.   

Nearly all active-duty personnel in the Fire Department of the City of New York (FDNY) responded to the World Trade Center (WTC) attacks on 9/11/2001.  These firefighters and emergency medical services (EMS) workers were likely exposed to noise and other ototoxic agents. Increased rates of self-reported hearing problems have been found among members of the WTC Health Registry, particularly if they were highly exposed to the WTC dust cloud. In this study of 8,646 FDNY personnel we evaluated changes in hearing over the decade after 9/11/2001 using audiometric records from regular occupational health exams.  In the first analysis, thresholds before the attack were compared with the first available results after the event to identify systematic trends in 15 dB threshold shifts across exposure level.  In the second analysis, we fit exponential time-to-event models to assess whether participants with greater exposures were at greater risk of 15 dB changes in threshold.  FDNY personnel who arrived at the WTC site on the morning of 9/11/2001 had greater odds of a threshold shift pre-9/11 to post-9/11, and personnel who arrived earlier and spent more time at the WTC site during the recovery effort were at greater risk of hearing loss.