Continuous Audio Recordings Research Articles

Sa1974 Characterizing Gastrointestinal (GI) Motility With a Computer-Aided, Non-Invasive Acoustic Sensor: Proof-of-Concept Testing in Normal Controls vs. Postoperative Patients

Background: motility disorders are highly prevalent and expensive. Motility testing is limited to invasive modalities, such as antroduodenal manometry or SmartPill, that provide one-time results not longitudinal tracking of dynamic physiology. Previous work showed that continuous audio recordings of bowel sounds from the abdominal wall correlates with manometry (Tomomasa AJG 1999). But until recently, there was no way to perform acoustic monitoring in a clinically feasible, cost-effective, point-of-care manner. Using advances in low-cost sensor technologies and computer-aided analysis, we developed a disposable, noninvasive Acoustic Surveillance (AGIS) sensor. We performed a proof-of-concept study to test whether monitoring can distinguish healthy controls (HC) from patients (Pts) recovering from colorectal surgery tolerating oral intake vs. those with postoperative ileus (POI): a model dysmotility syndrome suitable for initial testing. Methods: The device is a disposable plastic sensor embedded with a high-fidelity microphone that adheres to the abdominal wall (Figure). The device connects to a bedside gateway box that measures motility events with a signal-search engine. The results are displayed as GI telemetry and the computer calculates an AGIS index, defined as acoustic events scaled per second. We recruited HC subjects to wear the device for 2 hours following a breakfast of two eggs, a slice of bread, and a beverage. We compared their index to similar recordings of Pts recovering from colorectal surgery at a University-based VA medical center. We divided Pts into those tolerating postop feeding (Pts-Feed) vs. those with POI (Pts-POI) marked by nausea, vomiting, nasogastric lavage or failure to advance beyond sips. We used ANOVA to compare indices among groups and t-tests between groups. Results: There were 6 HC (mean age=28.5+3.1; 50%M), 7 Pts-Feed (age=65.4+4.6; 100%M) and 25 Pts-POI (age=63.4+10.7; 100%M). The mean+SD indices were 0.14+0.02, 0.03+0.0, and 0.016+0.01 events/second for HC, Pts-Feed and Pts-POI, respectively (ANOVA p<0.001). Using a diagnostic threshold of 0.1/second, the index was 97% sensitive and 85% specific for separating Pts from HC. When placed on a relative scale, the Pts-Feed and PtsPOI mean indices were only 26% and 11% that of HCs. Among postop Pts, there was a higher index in Pts-Feed vs Pts-POI (p=0.017). Conclusion: In this proof-ofconcept study, non-invasive monitoring of abdominal acoustics using a low-cost, disposable

Sleep Disordered Breathing in Children and Adolescents with Chiari Malformation Type I

Chiari malformation Type I (CM-I) has been associated with sleep disordered breathing (SDB). The aim of this study was to evaluate the prevalence of SDB in CM-I and its clinical correlates in a population of children and adolescents. Fifty-three consecutive children and adolescents affected by CM-I were enrolled (27 girls and 26 boys, mean age 10.3 ± 4.3, range: 3-18 years). All patients underwent neurological examination, MRI, and polysomnography (PSG). Otorhinolaryngologic clinical evaluation was performed in patients with polysomnographic evidence of sleep-related upper airway obstruction. Mean size of the herniation was 9.5 ± 5.4 mm. Fourteen patients had syringomyelia, 5 had hydrocephalus, 31 presented neurological signs, 14 had epileptic seizures, and 7 reported poor sleep. PSG revealed SDB in 13 subjects. Patients with SDB, compared to those without SDB, had a higher prevalence hydrocephalus (p = 0.002), syringomyelia (p = 0.001), and neurological symptoms (p = 0.028). No significant difference was observed in age, gender, prevalence of epilepsy, and size of the herniation. Obstructive SDB was associated with syringomyelia (p = 0.004), whereas central SDB was associated with hydrocephalus (p = 0.034). In our population of CM-I patients the prevalence of SDB was 24%, lower than that reported in literature. Moreover, our findings suggest that abnormalities in cerebrospinal fluid dynamics in CM-I, particularly syringomyelia and hydro-cephalus, are associated with SDB.

Obtaining calibrated sound pressure levels from consumer digital audio recorders

Audio recording of environmental sound is an increasingly efficient method for autonomously sensing many ecological and anthropogenic processes. The increasing capabilities of consumer digital audio recorders (DARs), especially increases in storage capacity and reductions in power consumption, enable continuous audio recordings exceeding 1month in duration with packages that are relatively small and inexpensive. To augment the ability of these systems to document the range of sounds present at a location, this paper examines two methods for calibrating recorders to measure sound levels. Compressed audio recorded by a DAR can be processed to yield relatively consistent measures of one-third octave band Leq values within a limited frequency and dynamic range. This was evaluated by synchronizing data with a Type-1 sound level meter. The calibration is stable over a 23day deployment outdoors with wide variation in ambient temperature and humidity. When considering aggregate acoustic metrics over time or a wide bandwidth such as an hourly A-weighted L50, the results can be quite accurate (within 1dBA).

Cultural soundscape of the Grant‐Kohrs Ranch national historic site.

It is often noted that although many parks and historic sites look as they did a century ago, very few sound as they once did. Management principles of U.S. National Parks and Historic Sites include the need to “preserve soundscape resources and values of the parks to the greatest extent possible to protect opportunities for appropriate transmission of cultural and historic sounds that are fundamental components of the purposes and values for which the parks were established.” Therefore, a critical part of soundscape management is to determine the existing ambient soundscape to detect trends and changes. An ongoing project at Grant‐Kohrs Ranch National Historic Site, located just north of Deer Lodge, Montana, involves a year‐long continuous audio recording and soundscape characterization. The recording status and example soundscape excerpts are presented and discussed.

Segmentation, Indexing, and Retrieval for Environmental and Natural Sounds

We propose a method for characterizing sound activity in fixed spaces through segmentation, indexing, and retrieval of continuous audio recordings. Regarding segmentation, we present a dynamic Bayesian network (DBN) that jointly infers onsets and end times of the most prominent sound events in the space, along with an extension of the algorithm for covering large spaces with distributed microphone arrays. Each segmented sound event is indexed with a hidden Markov model (HMM) that models the distribution of example-based queries that a user would employ to retrieve the event (or similar events). In order to increase the efficiency of the retrieval search, we recursively apply a modified spectral clustering algorithm to group similar sound events based on the distance between their corresponding HMMs. We then conduct a formal user study to obtain the relevancy decisions necessary for evaluation of our retrieval algorithm on both automatically and manually segmented sound clips. Furthermore, our segmentation and retrieval algorithms are shown to be effective in both quiet indoor and noisy outdoor recording conditions.

Detection of Cough Signals in Continuous Audio Recordings Using Hidden Markov Models

Cough is a common symptom of many respiratory diseases. The evaluation of its intensity and frequency of occurrence could provide valuable clinical information in the assessment of patients with chronic cough. In this paper we propose the use of hidden Markov models (HMMs) to automatically detect cough sounds from continuous ambulatory recordings. The recording system consists of a digital sound recorder and a microphone attached to the patient's chest. The recognition algorithm follows a keyword-spotting approach, with cough sounds representing the keywords. It was trained on 821 min selected from 10 ambulatory recordings, including 2473 manually labeled cough events, and tested on a database of nine recordings from separate patients with a total recording time of 3060 min and comprising 2155 cough events. The average detection rate was 82% at a false alarm rate of seven events/h, when considering only events above an energy threshold relative to each recording's average energy. These results suggest that HMMs can be applied to the detection of cough sounds from ambulatory patients. A postprocessing stage to perform a more detailed analysis on the detected events is under development, and could allow the rejection of some of the incorrectly detected events.

Performance of police first responders in utilizing automated external defibrillation on victims of sudden cardiac arrest

Objective. Rates of resuscitation from cardiac arrest are directly tied to time to defibrillation. To maximize results, the first arriving care provider should be equipped and trained to defibrillate. This would include police in those systems where they serve this function; to date, no training program has been examined for effectiveness in this group. The purpose of this study was to evaluate a training program designed to train police first respondes in the use of an automated external defibrillator (AED). Methods. One hundred seventy police officers previously trained to the level of first responders underwent a four-hour course to teach incorporation of the AED in their practice. The evaluation of police performance was assessed by written tests prior to, immediately after, and six months post initial training. Actual field use was evaluated by using separate data collection forms filled out at the time of the resuscitation by both police and EMS providers. Each trip sheet was also reviewed. Cassette tapes from the AED were reviewed for continuous ECG tracings and audio recordings to validate and confirm the previous data. Results. One hundred twenty-eight police cases were reviewed. The officers performed with few errors in AED operation, with the only problem areas being incorrect airway management and delay in performance of CPR to use the AED to reanalyze a nonshockable rhythm. These results were compared with those of the only two other studies examining the performance of first responders, which were EMTs and firefighters. The police results compared favorably with, and in some instances exceeded, those results. Conclusion. Police first responders trained in the use of AEDs performed at a level equivalent or superior to that in other reported studies. Future training strategies should stress proper integration of airway and CPR skills.