In the Neglect Radio smartphone app, it is essential to generate a convincing auditory illusion of a sound source moving from right to left in patients with left-sided neglect (or left to right in patients with right-sided neglect). This effect can be achieved using two primary approaches, dynamic volume control of stereo channels and simulation of a virtual sound source using head-related transfer functions (HRTFs).

A total of 37 healthy adult volunteers were recruited through university mailing lists at the University of Bern (Switzerland). Inclusion criteria required normal hearing and fluency in German or English. Participation was anonymous and voluntary, and no compensation was provided.

The study required the use of stereo headphones. Before beginning, participants were presented with a mandatory checkbox confirming that they were wearing headphones; only those who confirmed were able to proceed. Participants were instructed to adjust their device volume to a comfortable, medium level before starting. Because the study was conducted online, absolute sound levels could not be standardized, but the relative levels across the 3 auditory conditions were kept identical in all stimuli.

Each participant listened to audio samples of 5 everyday sound sources (a car, a crying baby, footsteps, a singing child, and a dog barking). Each sound was presented in three formats: (1) static stereo (no motion), (2) dynamic stereo volume control simulating left-to-right movement at 10° per second, and (3) HRTF-based audio simulating movement at 10° per second. Stimuli were presented in randomized order, and participants were not informed of the rendering method. In total, each participant rated 15 files. Realism was rated on a 0 to 100 scale (0=completely unrealistic, 100=highly realistic) using an online questionnaire interface.

Because all participants rated all 3 auditory conditions, a repeated-measures ANOVA was conducted to compare realism ratings across conditions. The Mauchly test indicated a violation of sphericity; therefore, Greenhouse-Geisser corrections were applied. Post hoc pairwise comparisons with Bonferroni adjustment were conducted to test differences between HRTF, dynamic volume control, and static stereo.

The sample of 37 participants was chosen to align with a previous perceptual study in spatial audio research that included 20 to 40 participants and was sufficiently powered to detect medium-to-large within-subject effects [20]. The subsequent usability evaluation was conducted with 10 participants, following established usability engineering guidelines suggesting that 5 to 10 users are sufficient to identify the majority of usability issues [21-23].

To maximize patient motivation and adherence, the selection of the audio content is a relevant factor. Two possible types of content for the audio signal were considered. The first type was realistic soundscapes, including everyday auditory experiences such as footsteps moving from right to left or a car driving past. Such sounds align with real-world expectations, making them highly effective in evoking spatial perception. Studies in ecological acoustics [24] suggest that familiarity with natural sounds enhances spatial awareness and immersion. The second type was entertaining content, including music, entertainment programs, and informational broadcasts and recordings. While not inherently directional, these sounds can be spatialized to create the illusion of a sound source travelling from the right to the left side. Research on immersive audio [25] indicates that even abstract audio elements, when processed through spatial techniques, can evoke a sense of movement.

Both approaches offer valuable contributions to the Neglect Radio concept. Realistic sounds provide strong spatial cues grounded in everyday experiences, while entertaining content potentially enhances engagement and listener retention. Based on the report by Kaufmann et al [13], we assume that repeatedly listening to realistic soundscapes alone, despite their significant spatial effect, would become somewhat monotonous over extended periods. Therefore, we suspect that effective everyday audio stimulation should contain entertaining content such as radio programs.

Listening habits among older adults support this choice. Surveys show that more than 75% of adults older than 30 years listen to radio weekly, and those over 60 often listen for several hours daily [26,27]. Many individuals older than 65 years continue to cherish personal music collections, often consisting of CDs or digital playlists featuring music that reflects the soundtrack of their youth. Research by Hays and Minichiello [28] suggests that listening to familiar music enhances cognitive function and emotional well-being in older adults. For practical and legal reasons, our app focuses on streaming publicly available web radio programs rather than personal music files, ensuring a wide variety of content without copyright barriers.

Based on the considerations described in the Creating the Illusion of a Moving Sound Source and Selecting Engaging Audio Content sections, as well as previous experiments [13-15], the Neglect Radio smartphone app was designed to fulfill the following functional requirements:

Based on functional requirement 6, we decided to use the Unity3D (version 5.0; Unity Technologies) programming environment for app development, allowing us to deploy the front-end app to both Android and iOS smartphones. Also, Unity3D provides a spatial audio function using the HRTF method. PHP (version 8.3) and SQL were used for the back end, including user management functions and recording of daily duration of use. Figure 1 shows the main user interface of the Neglect Radio app. The source code for the front end (in Unity3D) and the back end (in SQL and PHP) is available on Zenodo under an open-source license [29].

To assess the usability of the app, 10 participants in the spatial hearing evaluation (see the Creating the Illusion of a Moving Sound Source section) were invited to test a prototype of the app on a smartphone. The testing procedure involved a 15-minute radio-listening session through the app, followed by completion of an online version of the System Usability Scale (SUS) [30], originally developed by Brooke [31] for usability testing.

This study was conducted in accordance with the Declaration of Helsinki. The research protocol and its amendments were reviewed and approved by the Kantonale Ethikkommission Bern (2020-02464; approval granted July 8, 2021) and the Ethikkommission Nordwest- und Zentralschweiz. The amendment “Effect of visuo-tactile enhanced audio-kinetic stimulation in virtual reality for neglect patients” was approved under an expedited review, with TN serving as coordinating investigator and sponsor.

Participants did not receive financial compensation or other incentives for participation. Written informed consent was obtained from all participants (or their legally authorized representatives, where applicable) prior to any study procedures. Participants were informed that participation was voluntary and that they could withdraw at any time. Data used for analysis were deidentified before access by the research team. Direct identifiers (eg, names, full dates of birth, and contact details) were removed and only the given scores for algorithms and the results of the SUS were kept.

A total of 37 volunteers participated in the online study, listening to 5 audio sources in 3 different conditions each (HRTF, nonmoving stereo, and volume control) and rating them with respect to the “impression that the sound source moves from the right to the left.” The evaluation was conducted on a scale from 0 to 100, with 0 indicating a very unrealistic and 100 a very realistic impression. Detailed descriptive statistics for each condition are presented in Table 1. Figure 2 illustrates the distribution of realism ratings across the 3 audio conditions.

The realism rating for the HRTF-based rendering was significantly greater than for the volume control rendering (independent samples 1-tailed t test: t₇₂=3.722; P<.001), and the rating for the nonmoving stereo condition was significantly lower than for the other two groups (P<.001). A repeated-measures ANOVA with Greenhouse-Geisser correction showed a significant main effect of condition on realism ratings (F_1.43,52.36=120.6, P<.001; F_{1.43, 52.36}=120.6, P<.001; F_1.43,52.36=120.6, P<.001; Table 2). Mean realism ratings were highest for HRTF-based rendering (mean 80.9, SD 21.2), followed by dynamic volume control (mean 65.1, SD 14.6), and lowest for nonmoving stereo (mean 15.2, SD 16.4). Post hoc pairwise comparisons with Bonferroni correction confirmed that HRTF was rated significantly higher than volume control (P<.001), and both motion conditions (HRTF and volume control) were rated significantly higher than static stereo (both P<.001).

Ten healthy volunteers tested the Neglect Radio smartphone app for 15 minutes in a laboratory-based preliminary evaluation focused on feasibility in healthy users prior to clinical testing with stroke patients with hemispatial neglect. The mean SUS score was 86.5 (SD 6.9), exceeding the 68-point threshold for above-average usability and a rating of “excellent.” [32]. The individual SUS scores are shown in Figure 3.

This study introduces Neglect Radio, a smartphone-based auditory motion stimulation app. We evaluated the realism of its sound and its usability with healthy participants. Realism ratings confirmed that HRTF-based rendering produced the strongest sense of spatial motion, outperforming simple volume control and static stereo. This finding suggests that leveraging HRTFs can create a compelling illusion of sound motion, likely because it reproduces the subtle timing and filtering cues that the brain expects from real-world sound sources [17,18]. In a rehabilitation context, higher perceptual realism may be important, since realistic spatial cues could engage patients’ attention more effectively and promote a stronger sense of motion [33]. The app also yielded an excellent mean SUS score of 86.5 (SD 6.9), corresponding to superior usability [34]. This indicates that users, even those unfamiliar with advanced technology, found the app intuitive, comfortable, and engaging to use. Such positive reception is encouraging, as it meets a relevant requirement for rehabilitation tools: if an intervention is not user-friendly, patients (especially older stroke survivors) are unlikely to adopt it in their daily routine [15]. Taken together, these results validate the core design choices of Neglect Radio: using HRTF-based spatial sound to maximize realism and delivering the intervention via a smartphone app that patients can easily operate.

Importantly, these results should be interpreted within the broader landscape of neglect rehabilitation. Traditional approaches have focused heavily on visual scanning training and visuospatial cueing as top-down methods, with mixed evidence for long-term benefits [35]. Although visual scanning remains a widely recommended strategy, reviews highlight its limitations when used alone and the need for complementary modalities [35,36]. Multisensory stimulation, particularly using a bottom-up approach (eg, prism adaptation, limb activation, or optokinetic stimulation), has been proposed to enhance treatment effects by engaging additional sensory channels [36,37].

In this context, Neglect Radio introduces auditory motion stimulation as a novel mobile health–based tool. Unlike purely visual methods, auditory cues do not require intact visual processing or continuous effortful scanning, which is often impaired by anosognosia or attentional deficits in patients with neglect [35,37]. By leveraging spatialized sound, auditory motion may reach patients who cannot reliably engage with visual-only rehabilitation tasks, thereby broadening accessibility. This aligns with evidence that multisensory stimulation can lead to stronger and longer-lasting improvements than unimodal training [36,37].

The possible clinical potential of the Neglect Radio app lies in its accessibility and scalability, providing auditory stimulation to patients irrespective of their location, even allowing home-based use. Its intuitive interface makes it suitable for older users, aligning with real-world usability needs. Integrating this technology into existing rehabilitation protocols could enhance therapeutic outcomes by offering continuous attention training, from inpatient to home-based settings. Furthermore, the passive logging of app use provides clinicians with valuable adherence data, enabling them to deliver personalized care.

Beyond the clinical rationale, delivering auditory motion stimulation through a smartphone offers clear scalability advantages compared with clinic-based visual or multimodal interventions. Other mobile apps for neglect, such as Neglect App and Negami, have primarily targeted visual exploration tasks [38,39]. Similarly, music- or rhythm-based rehabilitation apps (eg, GotRhythm) and cognitive telerehabilitation apps (Rehastart) have shown that smartphones can effectively deliver therapy and achieve high usability [40,41]. Against this backdrop, Neglect Radio contributes a unique auditory-focused approach, which may synergize with existing digital interventions to create richer, multimodal rehabilitation strategies.

The primary limitation is that no patients with stroke were included; findings are therefore restricted to feasibility in healthy volunteers. This limitation is partially addressed by the work by Geiser et al [14,15], who used an early prototype for clinical evaluation, confirming similar feasibility results in patients with stroke.

Also, exposure time was short (15 minutes) and the sample sizes were modest (n=37 for realism; n=10 for usability). While these are typical limitations in perceptual and usability research, they limit conclusions about long-term adherence or variability in patient populations. Finally, the online study could not fully standardize sound levels across devices, although relative levels were controlled.

In addition, the online testing environment prevented full control over listening devices, volume levels, and background noise, which may have influenced realism ratings. The modest sample sizes, while typical for feasibility work, limit generalizability, and long-term adherence could not be assessed. We did not collect qualitative feedback, which would have provided additional insight into user experience. Finally, the technical implementation is dependent on headphones, and different headphone models or device types may affect the accuracy of spatial rendering.

Future research should focus on clinical validation through controlled trials involving patients with neglect and assess long-term adherence, patient engagement, and functional outcomes. Additionally, refinement of content personalization, integration of multimodal stimuli (eg, augmented reality or vibrotactile cues), and enhanced app customization could further optimize patient outcomes. Longitudinal studies are also needed to assess sustained user engagement and therapeutic benefits.

Neglect Radio provides a technically feasible and user-friendly platform for auditory motion stimulation. While healthy participants judged the stimuli as realistic and the app as highly usable, these results do not imply clinical efficacy. Future controlled clinical trials that include patients with stroke are essential to determine therapeutic effectiveness, adherence over time, and integration into rehabilitation practice.