The current LUCID approach utilizes the VIBE app to help provide a mask for tinnitus, manage the distress related to tinnitus, and potentially dampen stress reactivity when used regularly due to the increase in parasympathetic activity. Previous findings have demonstrated that chronic tinnitus patients have elevated stress reactivity. There has also been evidence to suggest that regular music therapy interventions may enhance parasympathetic activity in the autonomic nervous system. In the past, LUCID has advanced approaches in Artificial Intelligence technology, psychology, and neuroscience to explore the potential of music to improve health and well-being. Their work to date has focused on mental health indications related to stress and anxiety. This current study differs from their previous projects because of its exploration of tinnitus rather than stress and anxiety. Their customer-facing developments have been focused on an application programming interface that has been integrated by digital-health partners and a consumer-facing app (VIBE) that has been used to support research and development. The VIBE app (LUCID) incorporates theta-band (4 Hz) auditory beat stimulation and an auditory music recommendation system. The additive effect of these elements in the treatment of acute anxiety in individuals living with moderate trait anxiety was recently demonstrated in a randomized clinical trial.
The automated music recommendation system developed by LUCID employs the iso principle along with affective classification and reinforcement learning to cultivate affect-driven personalized music sequences. The iso principal is a methodology used in music therapy to achieve mood induction that involves matching musical stimuli to a patient's current mood and gradually changing the music in the direction of their desired mood state. The iso principle has been indicated in prior research to be more effective than other musical sequences at reducing tension. The system requires that a user input their current mood using the arousal and valence dimensions of the Russell Circumplex Model. Based on this input as well as the target emotional state of calm, the machine learning algorithm within the application predicts the optimal sequence of tracks to produce mood induction in the listener from their current emotional state to the target state. This machine learning algorithm uses reinforcement learning techniques and is trained on real-world data correlating the quantitative features of musical excerpts and sequences alongside the emotional responses induced by them in listeners.
For this study. there will be two groups with 25 participants per group. For the control condition, a white-noise treatment for the noise condition will be administered. The researchers will be using an open-label randomized controlled trial study where the participants are told that the researchers are attempting to assess the effectiveness of two common interventions for tinnitus. One approach involves broad-band masking with noise (Noise Condition), while the other uses music (LUCID Condition). Implementation of the noise condition will mirror the LUCID condition in terms of ease of access, look, feel, so that one condition does not look less professional than the other. Both conditions will be administered through the same app, and only the sound conditions will differ (white noise vs. LUCID music). All participants will be exposed to both the treatment and control conditions with the order of conditions counter-balanced (i.e., a cross-over design). The LUCID treatment will be compared to the active control across 50 people. The participant is expected to engage in daily use of VIBE app at a specified time (e.g., for 24 minutes daily) and as needed to help mask symptoms of tinnitus. To minimize variability in dosage across groups, the investigators will impose daily limits and minimum adherence standards. Participants will be permitted to use the app for a maximum of 48 minutes per day (i.e., 200% of recommended dosage). Participants who use the app greater than 60 minutes (250% of recommended dosage) or less than 12 minutes per day on average (50% of recommended dosage) or who use the app on less than 75% of the trial days will be considered non-adherent and will be removed from the final sample subjected to analyses. These types of adherence boundaries are typical for decentralized clinical trials involving digital therapeutics. The VIBE app will be deployed on the participant's smartphone, and they will be instructed to use their own headphones (i.e., there is no standardization of the playback system). The app has built-in features which allow us to assess adherence to treatment. Participants in the study will experience each intervention condition for a period of 4 weeks. It is encouraged that the participants stop the study if they believe the sound therapy is exacerbating their tinnitus, their reaction to tinnitus, or negatively affecting their hearing ability.
