A pilot test of sound-guided VR navigation
The paper asks whether assistive directional audio can lower the mental effort users spend navigating a virtual environment. Participants wore an HTC VIVE Focus Vision headset with a 100-degree field of view and stood on a BTrackS balance plate sampling at 25 Hz while a harness protected them from falls. When a user reached a path decision, the system played a sound localized to the left ear, right ear, or front to indicate turn left, turn right, or move forward.
The comparison that matters is against a no-feedback condition. In that baseline, participants wore the same equipment and followed the same procedure but received no additional audio, so they had to rely on visual scanning alone. The study's own framing is that this visual-only search raises mental workload, which is the claim the audio conditions were tested against.
Only six adults aged 18 and older took part (four male, two female; mean age 27 years). That sample size is the single most important qualifier on every result below: the paper calls the work a pilot study and states that a larger user study is needed to validate the preliminary findings.
Four audio categories, one shared navigation task
The study used 15 directional audio files grouped into four categories. Classical music included four Western pieces: Albinoni's Adagio in G minor, the Allegro con spirito from Mozart's Sonata K.448, Satie's Gymnopédie No. 1, and Beethoven's Für Elise. Natural soundscapes combined fountains, birdsong, and ocean waves. Electronic tracks were high-energy pieces with strong rhythmic drive and amplified bass. Unnatural sounds were human-made noises such as aviation, motorcycle, construction, and white noise.
These categories are not interchangeable background music. Each was chosen from prior work on attention, stress, or arousal, and the paper's hypothesis H4 predicted that classical and natural audio would produce significantly lower cognitive load than electronic and unnatural audio. Hypothesis H5 predicted a significant difference across all four categories.
Cognitive load itself was measured two ways: a NASA TLX self-report filled out after each task on a 0 to 10 scale, and two physiological proxies, center-of-pressure (CoP) velocity from the balance plate and pupil diameter from headset eye tracking. A 16-question Simulator Sickness Questionnaire was collected after the session to check cybersickness. Keeping these measures separate matters, because a self-report rating and a pupil measurement answer different questions about mental effort.
Classical and natural audio lowered reported load
The headline result is that auditory feedback significantly influenced cognitive load in the navigation task, with the abstract reporting p = .0011. The direction of the effect favored the calmer categories: classical and natural audio produced lower cognitive load than both the no-audio baseline and the other audio categories, which the authors read as support for H1 and H4.
The per-condition table is more cautious than that summary. Against no audio, classical showed a paired t of -2.47 with p = .0568 and a large effect size (d_z = -1.01); natural showed t = -2.33 with p = .0676 and d_z = -0.95. Neither crosses the conventional .05 threshold on its own. Electronic was weaker still (t = -1.40, p = .2209), and unnatural actually pointed the other way (t = 0.16, p = .875).
So the honest reading is a consistent direction of effect with large standardized differences for classical and natural audio, but individual pairwise tests that sit just above significance in a six-person sample. The significant overall p value and the non-significant pairwise values are not contradictory; they describe different comparisons, and readers should not treat the category ranking as firmly established.
Balance and pupil size track cognitive load
The paper's second contribution is the coupling between mental effort and body control. In the classical and natural conditions, better postural balance accompanied lower cognitive load. The authors explain this through dual-task interference: attention is a limited resource, so effort spent on navigation decisions leaves less for motor control, and VR balance is already stressed by sensory conflict and spatial uncertainty.
The correlation columns in the results table are where this claim lives, and they are mixed. For classical audio, the balance–load correlation was r = -0.82 with p = .0480, and the pupil–load correlation was r = 0.0236 with p reported as significant. For natural audio, balance–load reached r = -0.98 with p = .00048, while pupil–load was r = 0.803 with p = .0543. Electronic and unnatural conditions showed no significant correlations of the same kind.
Pupil diameter behaved as predicted in direction: higher cognitive load went with larger pupils in the calmer conditions, supporting H3. The paper describes this as pupil size reflecting the draw on cognitive resources. Because these are correlations within a small sample, they establish co-movement, not that audio caused balance to improve through reduced load.
Unnatural noise did not help and shifted balance
The unnatural category is the clearest negative case. Its cognitive-load comparison against no audio was essentially flat (t = 0.16, p = .875), so the human-made noise did not reduce reported effort. At the same time, the balance measure moved significantly in that condition (t = -2.75, p = .0401, d_z = -1.12), the only balance comparison in the table flagged as significant.
That combination is worth pausing on. A condition that fails to lower cognitive load but does shift postural measures suggests the noise was doing something to the body without relieving the mental task. The paper's design implication follows from this contrast: naturalistic, continuous, low-complexity sound is described as more comfortable and user-friendly than highly dynamic audio in assistive navigation.
The cybersickness questionnaire was included to check whether discomfort could explain load differences, but the packet does not report SSQ outcomes by condition, so no claim about cybersickness driving these results can be made from this evidence.
Where the design guidance reaches its limit
The authors frame auditory cues as an accessibility feature rather than a stylistic choice, arguing that designers should treat suitable audio feedback as fundamental, especially for older users or users with balance impairments. The reasoning is that users who are vulnerable to cognitive overload or balance difficulty gain the most from offloading path decisions onto sound.
That extension is a design argument, not a tested outcome. The participants were six adults with a mean age of 27, and no older or balance-impaired group was studied. The claim that the benefit would be larger for those users is plausible given the dual-task framing, but the paper provides no data from those populations, so it should be read as a hypothesis for the planned larger study.
The same caution applies to the category ranking. Classical and natural audio looked best on cognitive load, pupil diameter, and balance, but the pairwise significance is fragile at this sample size, and the electronic and unnatural categories were represented by only a handful of tracks each.
Prefer calm directional audio over dynamic noise
If this pilot is used to inform a VR navigation design decision, the actionable split is between calm and dynamic audio rather than between audio and silence alone. Classical and natural directional cues showed the largest reductions in reported cognitive load and the strongest accompanying balance and pupil correlations; electronic cues were weaker, and unnatural noise showed no load benefit while significantly shifting balance.
The priority the paper itself sets is to treat low-complexity, naturalistic directional sound as the default candidate for assistive navigation, then confirm it in a larger sample before writing it into accessibility standards. The six-participant design and the pairwise p values hovering near .05 mean the category ordering is a lead to test, not a threshold to design against.
Sources used in this answer
C-LISTEN: Cognitive Load Impacts of Sensory-Triggered Environmental Navigation in Virtual Reality
In a six-participant VR pilot, directional auditory cues significantly influenced cognitive load overall (p = .0011), with classical and natural audio showing the largest reductions versus no audio, alongside correlations between cognitive load, postural balance, and pupil diameter.