Natural sound may be intrinsically restorative regardless of its source attributions: Evidence from a running-river sound

Abstract

It remains unclear whether the restorative effects of listening to natural sounds reflect the effects of the sounds themselves or the influence of semantic labeling. This study examined whether manipulating sound-source attributions alters the self-reported restoration outcomes of listening to a 10-min running-river sound, as measured by the Restoration Outcome Scale (ROS). We hypothesized that listening itself would increase restoration outcomes without any source label (H1), that a “live-streamed from a national park” label would enhance the increase described in H1 (H2), and that an “artificially synthesized” label would attenuate the increase described in H1 (H3). In an online experiment, 168 subjects were randomly assigned to the “no-information (control),” “live-streamed,” or “synthesized” label groups. ROS was measured at baseline (T0), post-stress-induction/pre-listening (T1), and post-listening (T2); the labels were presented after the T1 ROS assessment, i.e., immediately before the listening. A linear mixed-effects model supported H1 but not H2–H3: restoration outcomes increased similarly across groups (no label effects). A post hoc qualitative analysis of the participants’ descriptions in their own words indicated that the sound evoked visual imagery of natural environments even without source information. These findings suggest that perceptual clarity of the sound source may itself act as an inherent label, indicating that bottom-up (stimulus-driven) and top-down (cognitive-driven) processes may be inseparable when listening to easily identifiable natural sound.