In Response: What Happens When Algorithmic Music Meets Pain Medicine

Dear Editors, We would like to thank Dr. Raglio for writing his letter in response to our original research article, “The Impact of Music on Nociceptive Processing,” and acknowledge several important points that he raises therein [1]. We very much agree that the analgesic effect of music is complex, with any given type of music being variably effective, based on individual preferences. This personal and interindividually variable nature of music enjoyment may in fact complicate the study of specific components of music impact on pain, making the discernment of relevant patient individual psychosocial phenotypes crucial to understanding variations in effect. While one can understand the average effect of certain song components (sound, melody, harmony, rhythm, texture, structure, or expression), any given element may impact individual listeners differently. To account for individualized thrills associated with music, some investigators have allowed study participants to select their preferred music from a predefined library or the participant’s personal library. Dr. Raglio’s letter stated that we had used such an approach, allowing subjects to choose their own music through an app. However, this actually was not the case, and we apologize if this was not clear from the methods. In our study, we in fact utilized an app called Unwind, which creates music algorithmically, but with several specific tracks available as samples for participants to hear, and participants then selected the one that appealed to them the most [2]. We read with great interest Dr. Raglio’s reference to Melomics-Health’s use of a generative algorithm to create new music based on a therapeutic logic [3]. As Dr. Raglio enumerates, this algorithmic strategy of music generation, which appears somewhat similar to the app we used, has several advantages to user-selected or therapist-selected music. Algorithmic music that employs isolated components of music that are known to be rewarding may present specific, culturally agnostic stimuli to participants in the laboratory setting, allowing a relatively proven but somewhat standardized stimulus to study the impact on pain mechanisms. Additionally, algorithmic music can produce “new” musical signatures, and as such it may add novelty, which is in itself engaging and rewarding, by remixing targeted interventions that have an evidentiary basis on laboratory work using the same musical components. Future work investigating the relative importance of novelty, familiarity, and emotional attachment vs specific musical elements to therapeutic benefit for a given individual is needed, as are systems that allow patient personalization.

As clinicians, we also recognize the importance of studying musical components that can be integrated into an acceptable intervention for the real-world patients we manage in our clinical practice. While we describe here the mechanistic effects of music on laboratory-based standardized pain stimuli, ultimately these effects must be evaluated among individuals who are experiencing clinical pain, and many previous studies have suggested an analgesic benefit. We have recently published a feasibility study in a clinical environment (emergency department), and we are beginning to understand how and whether and especially for which patients music interventions may be most helpful in clinical practice [4]. In particular, we found that patients with high levels of pain catastrophizing benefitted most. We look forward to other future mechanistic and pragmatic studies that assess the adjunctive integration of music with other pain therapies.

 

Funding sources: PRC was funded by K23DA044874, R44DA051106, The Hans and Mavis Lopater Psychosocal Fund, e-ink corporation; KLS was funded by R35GM128691; EWB was funded by R01DA047236.

Conflicts of interest: None.