4pSC15 – Reading aloud in a clear speaking style may interfere with sentence recognition memory – Sandie Keerstock

Reading aloud in a clear speaking style may interfere with sentence recognition memory

Sandie Keerstock – keerstock@utexas.edu
Rajka Smiljanic – rajka@austin.utexas.edu
Department of Linguistics, The University of Texas at Austin
305 E 23rd Street, B5100, Austin, TX 78712

Popular version of paper 4pSC15
Presented Thursday afternoon, May 16, 2019
177th ASA Meeting, Louisville, KY

Can you improve your memory by speaking clearly? If, for example, you are rehearsing for a presentation, what speaking style will better enhance your memory of the material: reading aloud in a clear speaking style, or reciting the words casually, as if speaking with a friend?

When conversing with a non-native listener or someone with a hearing problem, talkers spontaneously switch to clear speech: they slow down, speak louder, use a wider pitch range, and hyper-articulate their words. Compared to more casual speech, clear speech enhances a listener’s ability to understand speech in a noisy environment. Listeners also better recognize previously heard sentences and recall what was said if the information was spoken clearly.

In this study, we set out to examine whether talkers, too, have better memory of what they said if they pronounced it clearly.

Figure 1. Illustration of the procedure of the recognition memory task.

In the training phase of the experiment, 60 native and 30 non-native English speakers were instructed to read aloud and memorize 60 sentences containing high-frequency words, such as “The hot sun warmed the ground,” as they were presented one by one on a screen. Each screen directed the subject with regard to speaking style, alternating between “clear” and “casual” every ten slides. During the test phase, they were asked to identify as “old” or “new” 120 sentences written on the screen one at a time: 60 they had read aloud in either style, and 60 they had not.

Figure 2. Average of d’ (discrimination sensitivity index) for native (n=60) and non-native English speakers (n=30) for sentences produced in clear (light blue) and casual (dark blue) speaking styles. Higher d’ scores denote enhanced accuracy during the recognition memory task. Error bars represent standard error.

Unexpectedly, both native and non-native talkers in this experiment showed enhanced recognition memory for sentences they read aloud in a casual style. Unlike in perception, where hearing clearly spoken sentences improved listeners’ memory, findings from the present study tend to indicate a memory cost when talkers themselves produced clear sentences. This asymmetry between the production and perception effect on memory may be related to the same underlying mechanism, namely the Effortfulness Hypothesis (McCoy et al. 2005). In perception, more cognitive resources are used during processing of more-difficult-to-understand casual speech and fewer resources remain available for storing information in memory. Conversely, cognitive resources may be more depleted during the production of hyper-articulated clear sentences, which could lead to poorer memory encoding. This study suggests that the benefit of clear speech may be limited to the retention of spoken information in long-term memory of listeners, but not talkers.

4aSP4 – Streaming Video through Biological Tissues using Ultrasonic Communication – Gizem Tabak

Streaming Video through Biological Tissues using Ultrasonic Communication

Gizem Tabak – tabak2@illinois.edu
Michael Oelze – oelze@illinois.edu
Andrew Singer – acsinger@illinois.edu
University of Illinois at Urbana-Champaign
306 N Wright St
Urbana, IL 61801

Popular version of paper 4aSP4
Presented Thursday morning, May 16, 2019
177th ASA Meeting, Louisville, KY

Researchers at the University of Illinois at Urbana-Champaign have developed a fast, wireless communication alternative that also has biomedical implications. Instead of using radio frequency (RF) to transmit signals, the team is using ultrasonic waves to send signals at high enough data rates to transmit video through animal or human tissue.

The team of electrical and computer engineering professors Andrew Singer and Michael Oelze and graduate researcher Gizem Tabak have achieved a transmission rate of 4 megabits per second through animal tissue with 2-mm transmitting devices. This rate is high enough to send high definition video (3 Mbps) and 15 times faster than that RF waves can currently deliver.

CAPTION: Figure 1 – Experimental setup for streaming at 4Mbps through 2” beef liver).

The team is using this approach for communicating with implanted medical devices, like those used to scan tissue in a patients’ gastrointestinal (GI) tract.

Currently one of two methods are used to image the GI tract. The first is video endoscopy, which involves inserting a long probe with a camera and light down the throat to take real-time video and send it to an attached computer. This method has limitations in that it cannot reach the midsection of the GI tract and is highly invasive.

The second method involves a patient swallowing a pill that contains a mini camera that can take images throughout the tract. After a day or so, the pill is retrieved, and the physician can extract the images. This method, however, is entirely offline, meaning there is no real-time interaction with the camera inside the patient.

A third option uses the camera pill approach but sends the images through RF waves, which are absorbed by the surrounding tissue. Due to safety regulations governing electromagnetic radiation, the transmitted signal power is limited, resulting in data rates of only 267 kilobits per second.

The Illinois team is proposing to use ultrasound, a method that has already proven safe for medical imaging, as a communication method. Having achieved data rates of 4 Mbps with this system through animal tissue, the team is translating the approach to operate in real-time for use in the human body.

Pairing this communication technology with the camera pill approach, the device not only could send real-time video, but also could be remotely controlled. For example, it might travel to specific areas and rotate to arbitrary orientations. It may even be possible to take tissue samples for biopsy, essentially replacing endoscopic procedures or surgeries through such mini-remote controlled robotic devices.

4APP28 – Listening to music with bionic ears: Identification of musical instruments and genres by cochlear implant listeners – Ying Hsiao

“Listening to music with bionic ears: Identification of musical instruments and genres by cochlear implant listeners”

Ying Hsiao – ying_y_hsiao@rush.edu
Chad Walker
Megan Hebb
Kelly Brown
Jasper Oh
Stanley Sheft
Valeriy Shafiro – Valeriy_Shafiro@rush.edu
Department of Communication Disorders and Sciences
Rush University
600 S Paulina St
Chicago, IL 60612, USA

Kara Vasil
Aaron Moberly
Department of Otolaryngology – Head & Neck Surgery
Ohio State University Wexner Medical Center
410 W 10th Ave
Columbus, OH 43210, USA

Popular version of paper 4APP28
Presented Thursday morning, May 16, 2019
177th ASA Meeting, Louisville, KY

For many people, music is an integral part of everyday life. We hear it everywhere: cars, offices, hallways, elevators, restaurants, and, of course, concert halls and peoples’ homes. It can often make our day more pleasant and enjoyable, but its ubiquity also makes it easy to take it for granted. But imagine if the music you heard around you sounded garbled and distorted. What if you could no longer tell apart different instruments that were being played, rhythms were no longer clear, and much of it sounded out of tune? This unfortunate experience is common for people with hearing loss who hear through cochlear implants, or CIs, the prosthetic devices that convert sounds around a person to electrical signals that are then delivered directly to the auditory nerve, bypassing the natural sensory organ of hearing – the inner ear. Although CIs have been highly effective in improving speech perception for people with severe to profound hearing loss, music perception has remained difficult and frustrating for people with CIs.

Audio 1.mp4, “Music processed with the cochlear implant simulator, AngelSim by Emily Shannon Fu Foundation”
[insert audio 1 here]

Audio 2.mp4, “Original version [“Take Five” by Francesco Muliedda is licensed under CC BY-NC-SA]”
[insert audio 2 here]

To find out how well CI listeners identify musical instruments and music genres, we used a version of a previously developed test – Appreciation of Music in Cochlear Implantees (AMICI). Unlike other tests that examine music perception in CI listeners using simple-structured musical stimuli to pinpoint specific perceptual challenges, AMICI takes a more synthetic approach and uses real-world musical pieces, which are acoustically more complex. Our findings confirmed that CI listeners indeed have considerable deficits in music perception. Participants with CIs correctly identify musical instruments only 69% of the time and musical genres 56% of the time, whereas their age-matched normal-hearing peers identified instruments and genres with 99% and 96% correct, respectively. The easiest instrument for CI listeners were drums, which were correctly identified 98% of the time. In contrast, the most difficult instrument was flute, with only 18% identification accuracy. Flute was more often, 77% of the time, confused with string instruments. Among the genres, identification of classical music was the easiest, reaching 83% correct, while Latin and rock/pop music were most difficult (41% correct). Remarkably, CI listeners’ abilities to identify musical instruments and genres correlated with their ability to identify common environmental sounds (such as dog barking, car horn) and also spoken sentences in noise. These results provide a foundation for future work that will focus on rehabilitation in music perception for CI listeners, so that music may sound pleasing and enjoyable to them once again, with possible additional benefits for speech and environmental sound perception.


4aPA – Using Sound Waves to Quantify Erupted Volumes and Directionality of Volcanic Explosions – Alexandra Iezzi

“Using Sound Waves to Quantify Erupted Volumes and Directionality of Volcanic Explosions”

Alexandra Iezzi – amiezzi@alaska.edu
Geophysical Institute, Alaska Volcano Observatory
University of Alaska Fairbanks
2156 Koyukuk Drive
Fairbanks, AK 99775

David Fee – dfee1@alaska.edu
Geophysical Institute, Alaska Volcano Observatory
University of Alaska Fairbanks
2156 Koyukuk Drive
Fairbanks, AK 99775

Popular version of paper 4aPA
Presented Thursday morning, May 16, 2019
177th ASA Meeting, Louisville, KY

Volcanic eruptions can produce serious hazards, including ash plumes, lava flows, pyroclastic flows, and lahars. Volcanic phenomena, especially explosions, produce a substantial amount of sound, particularly in the infrasound band (<20 Hz, below human hearing) that can be detected at both local and global distances using dedicated infrasound sensors. Recent research has focused on inverting infrasound data collected within a few kilometers of an explosion, which can provide robust estimates of the mass and volume of erupted material in near real time. While the backbone of local geophysical monitoring of volcanoes typically relies on seismometers, it can sometimes be difficult to determine whether a signal originates from the subsurface only or has become subaerial (i.e. erupting). Volcano infrasound recordings can be combined with seismic monitoring to help illuminate whether or not material is actually coming out of the volcano, therefore posing a potential threat to society.

This presentation aims to summarize results from many recent studies on acoustic source inversions for volcanoes, including a recent study by Iezzi et al. (in review) at Yasur volcano, Vanuatu. Yasur is easily accessible and has explosions every 1 to 4 minutes making it a great place to study volcanic explosion mechanisms (Video 1).

Video 1 [Iezzi VIDEO1.mp4] – Video of a typical explosion at Yasur volcano, Vanuatu.

Most volcano infrasound inversion studies assume that sound radiates equally in all directions. However, the potential for acoustic directionality from the volcano infrasound source mechanism is not well understood due to infrasound sensors usually being deployed only on Earth’s surface. In our study, we placed an infrasound sensor on a tethered balloon that was walked around the volcano to measure the acoustic wavefield above Earth’s surface and investigate possible acoustic directionality (Figure 1).

Figure 1 [balloon.JPG] – Image showing the aerostat on the ground prior to launch (left) and when tethered near the crater rim of Yasur (right).

Volcanos typically have high topographic relief that can significantly distort the waveform we record, even at distances of only a few kilometers. We can account for this effect by modeling the acoustic propagation over the topography (Video 2).

Video 2 [Iezzi VIDEO2.mp4] – Video showing the pressure field that results from inputting a simple compressional source at the volcanic vent and propagating the wavefield over a model of topography. The red denotes positive pressure (compression) and blue denotes negative pressure (rarefaction). We note that all complexity past the first red band is due to topography.

Once the effects of topography are constrained, we can assume that when we are very close to the source, all other complexity in the infrasound data is due to the acoustic source. This allows us to solve for the volume flow rate (potentially in real time). In addition, we can examine directionality for all explosions, which may lead to volcanic ejecta being launched more often and farther in one direction than in others. This poses a great hazard to tourists and locals near the volcano and may be mitigated by studying the acoustic source from a safe distance using infrasound.

4aBAa7 – Unprecedented high-spatial resolution was achieved in ultrasound imaging by breaking the fundamental limitation with the operating ultrasound wavelength – Kang Kim

Unprecedented high-spatial resolution was achieved in ultrasound imaging by breaking the fundamental limitation with the operating ultrasound wavelength

Kang Kim – kangkim@upmc.edu
Qiyang Chen – qic41@pitt.edu
Jaesok Yu – jaesok.yu@ece.gatech.edu
Roderick J Tan – tanrj@upmc.edu
University of Pittsburgh
3550 Terrace St, room 623, Pittsburgh, PA 15261

Popular version of paper 2aBA8; 4aBAa7
Presented Tuesday & Thursday morning, May 14 & 16, 2019
177th ASA Meeting, Louisville, KY

US imaging is one of the most favored imaging modalities in clinics in general because of its real-time display, safety, noninvasiveness, portability and affordability. One major disadvantage of ultrasound imaging is its limited spatial resolution that is fundamentally governed by the wavelength of the operating ultrasound. We developed a new super-resolution imaging algorithm that can achieve super high-spatial resolution beyond such limitation called acoustic diffraction limit.
The concept of the super resolution that bypasses a physical limit for the maximum resolution of traditional optical imaging was originally introduced in microscopy imaging community and later developed into a ground-breaking technology of the nano-dimension microscopy imaging, for which the Nobel Prize in Chemistry was awarded in 2014. In brief, microscopy super resolution imaging technology is based on randomly repeated blinking process of the fluorophores in response to the light source of the microscopy. In recent years, the concept has been translated into ultrasound imaging community. The random blinking process that requires for achieving super resolution using ultrasound is provided by flowing microbubbles in blood vessels which randomly oscillate in response to the ultrasound pressure from the imaging transducer. The maximum spatial resolution in super resolution microscopy technology is in the range of tens of nanometers (10-9 m) that allows to visualize the pathways of individual molecules inside living cells, while ultrasound super resolution imaging can achieve a spatial resolution in the range of tens of micrometers (10-6 m) when using a typical clinical ultrasound imaging transducer of a few MHz center frequency. However, due to the large imaging depth of ultrasound up to several centimeters, ultrasound super resolution imaging technology is practically very useful in imaging human subject with greater details of microvasculature which is of critical importance for many diseases.
Figure 1

Traditional contrast enhanced ultrasound (CEU) imaging technologies using microbubbles provide superior contrast of vasculatures, effectively suppressing the surrounding tissue signals, but the spatial resolution remains to the acoustic diffraction limit. In recent years, to overcome such limitation with CEU, several approaches have been made to overcome such limitation by employing super resolution concept, however requiring a long scan time, which hinders the technology from being wide spread. The major contribution from my laboratory is to drastically shorten the scan time of super resolution imaging using deconvolution algorithm for microbubble center localization, as well as to compensate artifacts due to physiological motions using block matching based motion correction and spatio-temporal-interframe-correlation based data re-alignment, so that the technology can be used in vivo for diverse applications. In brief, a novel approach of ultrafast ultrasound imaging, rigid motion compensation, tissue signal suppressor and deconvolution based deblurring has been developed for both high spatial and temporal resolution.

Video 1

The developed technology was applied in imaging microvasculature change which is a critical feature during disease development and progress. Vasa vasorum that is network of small blood vessels that supply the walls of large blood vessels and often multiplies and infiltrates into atherosclerotic plaque were identified in rabbit model.

Microvascular rarefaction is a key signature of acute kidney injury that often progress into chronic kidney diseases and eventual kidney failure. Microvessels in mouse acute kidney injury model were successfully identified and quantitatively analyzed.

Figure 3

4aAB1 – The best available science? Are NOAA Fisheries marine mammal exposure noise guidelines up to date? – Michael Stocker

The best available science? Are NOAA Fisheries marine mammal exposure noise guidelines up to date?

Michael Stocker – mstocker@OCR.org
Ocean Conservation Research
P.O. Box 559
Lagunitas, California 94938

Popular version of paper 4aAB1
Presented Thursday morning, May 16, 2019
177th ASA Meeting, Louisville, KY


NOAA Fisheries employs a set of in-water noise exposure guidelines that establish regulatory thresholds for ocean actions that impact marine mammals. These are established based on two impact criteria: Level A – a physiological impact, and Level B – a behavioral impact or disruption. Since the introduction of these exposure definitions, much more work has been published on behavioral impacts of various noise exposures, and consideration of other variables such as frequency, sound quality, and multiple sound-source exposures. But these variables have not yet been incorporated into the NOAA Fisheries exposure guidelines.

Determining regulatory thresholds

In the Marine Mammal Protection Act (MMPA) sound exposure levels are categorized in two levels, Level A” and “Level B.” “Level A Take” defined by the National Marine Fisheries Service (NMFS) as a “do not exceed” threshold below which physical injury would not occur. In whales and whales, dolphins, and porpoises this was 180dB (re: 1μPa).

A “Level B Take” is defined as “any act that disturbs or is likely to disturb a marine mammal or marine mammal stock in the wild by causing disruption of natural behavioral patterns, including, but not limited to, migration, surfacing, nursing, breeding, feeding, or sheltering, to a point where such behavioral patterns are abandoned or significantly altered.” But defining what constitutes “disruption” is fraught with threshold vagaries – given that behavior is always contextual, and the weight of the “biological significance” of the disruption hinges on a human value scale. How biologically significant is it when Bowhead whales change their vocalization rates in response to barely audible airgun exposure, well below the Level B threshold? How biologically significant is it when a sea lion risks exposure to loud, intentionally (above Level A) Acoustic Harassment Devices intended to scare sea lions away from fish farms actually attracts them by letting them know that “dinner” is available.

Regulatory Metrics

Regulations work best when they are unambiguous. Regulators are not fond of nuance. Dichotomous decisions of Yes/No, Go/No-Go are their stock and trade. It was for this reason that until just recently the marine mammal exposure guidelines were really simple:

Noise exposure above 180dB = Level A exposure.
Noise exposure above 160dB = Level B exposure (for impulsive sounds)
Noise exposure above 120dB = Level B exposure (for continuous sounds)

But it was clear that these original regulatory thresholds were actually too simple. When dolphins ride the bow waves of seismic survey vessels – frolicking in a Level A noise field, it was apparent that the regulatory thresholds did not reflect common field conditions. This was recently addressed in guidelines that more accurately reflected the noise exposure criteria relative to the hearing ranges of a range of the various marine mammal species – from large “Low Frequency” baleen whales, to small “High Frequency” dolphins and porpoises. While this new standard more accurately reflects the frequency-defined hearing ranges of the exposed animals, it does not accurately address the complexity of the noise exposures in terms of sound qualities, nor in terms of the complexity of the sound environments in which the exposures would typically occur.

Actual sound exposures

Increasingly complex signals are being used in the sea for underwater communication and equipment control. These communication signals can be rough or “screechy” sounding and more disturbing and more damaging than the simple signals used for auditory testing.

Additionally, when sounds presented in a typical Environmental Impact Statements, they are presented as single sources of sound. And while there is some consideration for accumulated noise impacts, the accumulation period “resets” after 24 hours, so the metric only reflects accumulated noise exposure and does not address the impacts of a habitat completely transformed by continuous, or ongoing noise. Given that typical seismic airgun surveys run around the clock for weeks to months at a time, and have an acoustical reach of hundreds to thousands of kilometers, the activity is likely to have much greater behavioral impact than is reflected in accumulating and dumping of a noise exposure index every 24 hours.

Furthermore, operations such as seismic survey, or underwater extraction industry operations typically use a lot of different, but simultaneous sound sources. Seismic surveys may include seafloor profiling with multi-beam or side-scan sonars. Underwater extraction industries such as seafloor processing for oil and gas extraction, or seafloor mining operations will necessarily have multiple sound sources – with noisy equipment, along with acoustical communications for status monitoring, and acoustical remote control of the equipment. These concurrently operating compliments of equipment can create a very complex soundscape. And even if the specific pieces of equipment don’t in-and-of-themselves exceed regulatory thresholds, they may nonetheless create acoustically-hostile soundscapes likely to have behavioral and metabolic impacts on marine animals. So far there is no qualitative metrics for compromised soundscapes, but modeling for concurrent sound exposures is possible, and in this context, many concurrent sounds would constitute “continuous sound,” thereby qualifying the soundscape as a whole under the Level B continuous sound criteria of 120dB.

This is particularly the case for a proposed set of seismic surveys in the Mid-Atlantic, wherein three separate geophysical surveys would be occurring simultaneously in close proximity. “Incidental Harassment Authorizations” have been released by NOAA Fisheries for these surveys which have not taken the ‘concurrent noise exposures’ into account.

Additionally, while sound sources in the near-field may be considered “impulsive sounds.” And thus regulated under “Level B” criteria for impulse sounds, due to reverberation, louder sounds which have a long reach should be considered as “continuous sound sources” and thus be regulated under the Level B ‘continuous sound’ criteria of 120dB.


1. NOAA sound exposure metric should be updated to reflect sound quality (accommodating for signal characteristics) as well as amplitude.
2. “Soundscapes” need qualitative and quantitative definitions, and then incorporated into the regulatory framework.
3. Exposure metrics needs to accommodate for concurrent sound source exposures.
4. The threshold for what constitutes “continuous sound” needs to be more clearly defined, particularly in terms of loud sound sources in the far field subject to reverberation and “multi-path” echoes.