Sound recognition is effortless for humans but poses a significant challenge for artificial hearing systems. Deep neural networks (DNNs), especially convolutional neural networks (CNNs), have recently surpassed traditional machine learning in sound classification. However, current DNNs map sounds to labels using binary categorical variables, neglecting the semantic relations between labels. Cognitive neuroscience research suggests that human listeners exploit such semantic information besides acoustic cues. Hence, our hypothesis is that incorporating semantic information improves DNN\'s sound recognition performance, emulating human behaviour. In our approach, sound recognition is framed as a regression problem, with CNNs trained to map spectrograms to continuous semantic representations from NLP models (Word2Vec, BERT, and CLAP text encoder). Two DNN types were trained: semDNN with continuous embeddings and catDNN with categorical labels, both with a dataset extracted from a collection of 388,211 sounds enriched with semantic descriptions. Evaluations across four external datasets, confirmed the superiority of semantic labeling from semDNN compared to catDNN, preserving higher-level relations. Importantly, an analysis of human similarity ratings for natural sounds, showed that semDNN approximated human listener behaviour better than catDNN, other DNNs, and NLP models. Our work contributes to understanding the role of semantics in sound recognition, bridging the gap between artificial systems and human auditory perception.

Restoration of vulnerable marine habitats is becoming increasingly popular to cope with widespread habitat loss and the resulting decline in biodiversity and ecosystem services. Lately, restoration strategies have been employed to enhance the recovery of degraded meadows of the Mediterranean endemic seagrass Posidonia oceanica. Typically, habitat restoration success is evaluated by the persistence of foundation species after transplantation (e.g., plant survival and growth) on the short and long-term, although successful plant responses do not necessarily reflect the recovery of ecosystem biodiversity and functions. Recently, soundscape (the spatial, temporal and frequency attribute of ambient sound and types of sound sources characterizing it) has been related to different habitat conditions and community structures. Thus, a successful restoration action should lead to acoustic restoration and soundscape ecology could represent an important component of restoration monitoring, leading to assess successful habitat and community restoration. Here, we evaluated acoustic community and metrics in a P. oceanica restored meadow and tested whether the plant transplant effectiveness after one year was accompanied by a restored soundscape. With this goal, acoustic recordings from degraded, transplanted and reference meadows were collected in Sardinia (Italy) using passive acoustic monitoring devices. Soundscape at each meadow type was examined using both spectral analysis and classification of fish calls based on a catalogue of fish sounds from the Mediterranean Sea. Seven different fish sounds were recorded: most of them were present in the reference and transplanted meadows and were associated to Sciaena umbra and Scorpaena spp. Sound Pressure Level (SPL, in dB re: 1 μPa-rms) and Acoustic Complexity Index (ACI) were influenced by the meadow type. Particularly higher values were associated to the transplanted meadow. SPL and ACI calculated in the 200-2000 Hz frequency band were also related to high abundance of fish sounds (chorus). These results showed that meadow restoration may lead to the recovery of soundscape and the associated community, suggesting that short term acoustic monitoring can provide complementary information to evaluate seagrass restoration success.

BACKGROUND: Diabetic retinopathy (DR), a chronic and progressive microvascular complication of diabetes mellitus, substantially threatens vision and is a leading cause of blindness among working-age individuals worldwide. Traditional diagnostic methods, such as ophthalmoscopy and fluorescein angiography are nonquantitative, invasive, and time consuming. Analysis of protein biomarkers in tear fluid offers noninvasive insights into ocular and systemic health, aiding in early DR detection. This study introduces a surface acoustic wave (SAW) microchip that rapidly enhances fluorescence in bead-based immunoassays for the sensitive and noninvasive DR detection from human tear samples.
RESULTS: The device facilitated particle mixing for immunoassay formation and particle concentration in the droplet, resulting in an enhanced immunofluorescence signal. This detachable SAW microchip allows the disposal of the cover glass after every use, thereby improving the reusability of the interdigital transducer and minimizing potential cross-contamination. A preliminary clinical test was conducted on a cohort of 10 volunteers, including DR patients and healthy individuals. The results demonstrated strong agreement with ELISA studies, validating the high accuracy rate of the SAW microchip.
CONCLUSIONS: This comprehensive study offers significant insights into the potential application of a novel SAW microchip for the early detection of DR in individuals with diabetes. By utilizing protein biomarkers found in tear fluid, the device facilitates noninvasive, rapid, and sensitive detection, potentially revolutionizing DR diagnostics and improving patient outcomes through timely intervention and management of this vision-threatening condition.

Audible very-high frequency sound (VHFS) and ultrasound (US) have been rated more unpleasant than lower frequency sounds when presented to listeners at similar sensation levels (SLs). In this study, 17 participants rated the sensory unpleasantness of 14-, 16-, and 18-kHz tones and a 1-kHz reference tone. Tones were presented at equal subjective loudness levels for each individual, corresponding to levels of 10, 20, and 30 dB SL measured at 1 kHz. Participants were categorized as either \"symptomatic\" or \"asymptomatic\" based on self-reported previous symptoms that they attributed to exposure to VHFS/US. In both groups, subjective loudness increased more rapidly with sound pressure level for VHFS/US than for the 1-kHz reference tone, which is consistent with a reduced dynamic range at the higher frequencies. For loudness-matched tones, participants rated VHFS/US as more unpleasant than that for the 1-kHz reference. These results suggest that increased sensory unpleasantness and reduced dynamic range at high frequencies should be considered when designing or deploying equipment which emits VHFS/US that could be audible to exposed people.

The study of the perceived affective qualities (PAQs) in soundscape assessments have increased in recent years, with methods varying from in-situ to laboratory. Through technological advances, virtual reality (VR) has facilitated evaluations of multiple locations in the same experiment. In this paper, VR reproductions of different urban sites were presented in an online and laboratory environment testing three locations in Greater Manchester (\'Park\', \'Plaza\', and pedestrian \'Street\') in two population densities (empty and busy) using ISO/TS 12913-2 (2018) soundscape PAQs. The studied areas had audio and video recordings prepared for 360 video and binaural audio VR reproductions. The aims were to observe population density effects within locations (Wilcoxon test) and variations between locations (Mann-Whitney U test) within methods. Population density and comparisons among locations demonstrated a significant effect on most PAQs. Results also suggested that big cities can present homogenous sounds, composing a \'blended\' urban soundscape, independently of functionality. These findings can support urban design in a low-cost approach, where urban planners can test different scenarios and interventions.

Turbulence and sound are important cues for oyster reef larval recruitment. Numerous studies have found a relationship between turbulence intensity and swimming behaviors of marine larvae, while others have documented the importance of sounds in enhancing larval recruitment to oyster reefs. However, the relationship between turbulence and the reef soundscape is not well understood. In this study we made side-by-side acoustic Doppler velocimeter turbulence measurements and hydrophone soundscape recordings over 2 intertidal oyster reefs (1 natural and 1 restored) and 1 adjacent bare mudflat as a reference. Sound pressure levels (SPL) were similar across all three sites, although SPL > 2000 Hz was highest at the restored reef, likely due to its larger area that contained a greater number of sound-producing organisms. Flow noise (FN), defined as the mean of pressure fluctuations recorded by the hydrophone at f < 100 Hz, was significantly related to mean flow speed, turbulent kinetic energy, and turbulence dissipation rate (ε), agreeing with theoretical calculations for turbulence. Our results also show a similar relationship between ε and FN to what has been previously reported for ε vs. downward larval swimming velocity (w b ), with both FN and w b demonstrating rapid growth at ε > 0.1 cm2 s-3. These results suggest that reef turbulence and sounds may attract oyster larvae in complementary and synergistic ways.

This study examined the influence of stimulus properties on sound externalization when listening with hearing aids. Normally hearing listeners were presented with broadband \"tokens\" (environmental sounds and speech) from loudspeakers, and rated externalization using a continuous scale. In separate blocks, they listened unaided or while wearing behind-the-ear hearing aids with closed domes and low gain (linear or compressive). There was a significant influence of token on ratings, even for unaided listening, and the effect of hearing aids depended on token. An acoustic analysis indicated that hearing aids were more likely to disrupt externalization for peakier sounds with a low-frequency emphasis.

The focus of this perspective paper is on relationships between sound-producing body motion and corresponding perceived sound features, guided by the idea of shapes as the common denominator of these two domains. The term shape is used to denote graphical-pictorial renderings of phenomena that we perceive or imagine, and may have physical manifestations as tracings on paper or on screen, or as gesticulations, or just as imagined tracings in our minds. Shapes give us intermittent snapshots of unfolding motion and sound fragments, and the point of shapes is to make ephemeral sound and motion features tractable as more permanent objects. Shapes of perceived sound include dynamic, spectral, textural, pitch-related, harmonic, etc. features as shapes, whereas shapes of sound-producing motion include both motion trajectories and postures of sound-producing effectors, i.e., of fingers, hands, arms, etc., or mouth, lips, and tongue.

The rapid advancement of AI and machine learning has significantly enhanced sound and acoustic recognition technologies, moving beyond traditional models to more sophisticated neural network-based methods. Among these, Spiking Neural Networks (SNNs) are particularly noteworthy. SNNs mimic biological neurons and operate on principles similar to the human brain, using analog computing mechanisms. This capability allows for efficient sound processing with low power consumption and minimal latency, ideal for real-time applications in embedded systems. This paper reviews recent developments in SNNs for sound recognition, underscoring their potential to overcome the limitations of digital computing and suggesting directions for future research. The unique attributes of SNNs could lead to breakthroughs in mimicking human auditory processing more closely.

Unpredictable deviations from an otherwise regular auditory sequence, as well as rare sounds following a period of silence, are detected automatically. Recent evidence suggests that the latter also elicit quick involuntary modulations of ongoing motor activity emerging as early as 100 ms following sound onset, which was attributed to supramodal processing. We explored such force modulations for both rare and deviant sounds. Participants (N = 29) pinched a force sensitive device and maintained a force of 1-2 N for periods of 1 min. Task-irrelevant tones were presented under two conditions. In the Rare condition, 4000 Hz tones were presented every 8-to-16 s. In the Roving condition, 4000 Hz and 2996 Hz tones were presented at rate of 1 s, with infrequent (p = 1/12) frequency changes. In the Rare condition, transient force modulations were observed with a significant increase at ~ 234 ms, and a decrease at ~ 350 ms. In the Roving condition with low frequency deviant tones, an increase in force was observed at ~ 277 ms followed by a decrease at ~ 413 ms. No significant modulations were observed during perception of high frequency deviants. These results suggest that both rare silence-breaking sounds and low-pitched deviants evoke automatic fluctuations of motor responses, which opens up the possibility that these force modulations are triggered by stimulus-specific change-detection processes.