Although cognitive loss is the hallmark of dementia, and language deficits are frequently observed (Kemplar & Goral, 2008), certain music skills seem to remain intact, such as the ability to sing familiar songs, the musical details (i.e., melody, rhythm or lyrics), and the embedded memories and rich associations related to the song (Clair, 2000; Tomaino, 2002; Vanstone & Cuddy, 2020). Singing these songs can trigger long-term memories and feelings associated with the people, places and conditions that were present when these songs were learned or listened to (Vink & Hanser, 2018). These learned associations create physiological and emotional responses and can enrich one’s sense of self in place and time (Baird & Thompson, 2018; Clements-Cortes & Bartel, 2018). For people with dementia, it can promote relatedness and security and have a positive effect on mood and emotions (Clements-Cortes & Bartel, 2018).

Preferred music can be a valuable tool in assisting people with dementia to connect with their identity and life history, acting as a springboard for reminiscence (McDermott et al., 2013). One of the most robust findings in autobiographical memory research is the concept of the ‘reminiscence bump’ (Rubin et al., 1986). This is the tendency for people to more readily access memories from childhood to about 30 years of age (Koppel & Rublin, 2016; Sherman, 2013). These decades of a person’s life are a particularly rich period, which evoke increased music-related memories (Baird et al. 2018). In a communal setting, music’s memory-inducing effects may give rise to the recovery of one's narrative agency, allowing both the caregiver and person with dementia to interact in a more mutually engaging way and create a meaningful social connection (Matthews, 2015).

There is an accumulating body of evidence that suggests music therapy is a creative, non-pharmacological, and a cost-effective intervention for older adults (Aldridge, 2000; Baird, Garrido & Tamplin, 2020; Belgrave et al., 2011; Bright, 1997a, 1997b; Clair and Memmott, 2008; Hanser et al., 2020; Tamplin & Clark, 2020). For people with dementia, it provides a person-centred and supportive intervention that can bolster their well-being by promoting social engagement, communication, and improved quality of life. Its primary therapeutic aim is to meet psychosocial needs and to understand the person with dementia on an emotional level (Ridder & Wheeler, 2015). Other goal areas have included: improving and maintaining cognitive, emotional, and physical skills, and reducing BPSD such as restlessness or agitation (Aldridge, 2000; Ho et al., 2019; Ledger & Baker, 2007).

There are a variety of interventions used by music therapists in dementia care, with an emphasis on the therapeutic relationship and process (Cho, 2018). Individual music therapy including singing, reminiscence, and therapeutic songwriting have all been employed (Ahessy, 2017; Baker, 2015; Baker & Stretton-Smith, 2017; McDermott et al., 2013; Tamplin & Clark, 2020). Group interventions have traditionally included: group music therapy, improvisation, drumming, and singing (Lesta & Petocz, 2006; McDermott et al., 2013). In recent years there has been an emergence of therapeutic choir work (Ahessy, 2016; Cho, 2018; Clements-Cortez, 2014, 2015; Tamplin & Clark, 2020), and community group singing has been used to support people with dementia as well as their caregivers (Clark et al., 2018; Tamplin et al., 2018).

When working with people with dementia, reminiscence often organically occurs as a result of singing familiar songs. This has been documented anecdotally but has not been extensively studied. Ashida (2000) investigated the effect of a 5-day group reminiscence music therapy intervention on depressive symptoms in people with dementia. Familiar songs were presented related to the theme of the day, with accompanying questions relating to the lyrical content, enabling conversation, and social interaction. A significant decrease in depressive symptoms was observed. Dassa & Amir (2014) investigated the role of singing familiar songs in encouraging conversation among this population. Songs from the participants’ past elicited memories related to their social and national identity, encouraging members to contribute spontaneously.

Five studies combining music and reminiscence therapy interventions in elderly populations were identified in a systematic review by Istvandity (2017). All of the identified studies used familiar music to stimulate reminiscence activities. Four studies had a positive effect for participants (Ashida, 2000; Mohammadi, 2011; Rawtaer et. al, 2015; Takahashi & Matsushita, 2006), but lacked consistency, apart from their use of singing of familiar songs. Furthermore, the structure, delivery method, and evaluation of the delivery method were not reported. The review demonstrates the predominantly positive effects of combined music and reminiscence interventions upon the mental well-being of participants, especially stress, anxiety, and depression. However, there remains a lack of research on the effectiveness of this intervention to promote emotional and social well-being (Istvandity, 2017).

Webster et al. (2010) propose three basic types of reminiscence: simple or unstructured reminiscence, structured reminiscence, or life review and life-review therapy.

Reminiscence work, whether prompted or structured, generally involves group sessions, and is an appropriate tool for linking people with dementia who may be isolated back into a wider social fabric (Schweitzer, 2011). Reminiscence therapy and reminiscence work do not have a large evidence base and require increased rigor in effectiveness studies (Woods et al., 2005). Despite these limitations, they are a widely used form of engagement between staff and people with dementia in residential care settings (DSDC, 2002). Group-based reminiscence therapy has demonstrated positive effects on cognition (Wang, 2007), reduction of depressive symptoms (Chiang et al., 2010), and promotion of positive self-esteem (Chao et al., 2006). The three-Cs model (Bender et al., 1999), highlights the benefits of reminiscence “for clients, such as interacting, socialising, learning and engaging in therapeutic activities; benefits for carers to aid communication and improve staff skills, and benefits for the work context or culture of the unit” (Parker, 2006, p. 45). By creatively linking the past with the present, the resulting sense of continuity may restore a sense of personal identity, mastery, self-esteem, and integrity that may be otherwise lost (Gibson, 2006; Kitwood, 1997; Schweitzer & Bruce, 2008). This innovative approach to reminiscence has enabled many people with dementia to overcome their sense of isolation and become valued members of ongoing social groups (Schweitzer, 2011).

There is a lack of conceptual clarity surrounding reminiscence, thus leading to poor consistency of terminology and taxonomy in the field (Webster et al., 2010). Reminiscence, recollection, life stories, chronological life review, reminiscence therapy, and narrative therapy are some of the terms used to describe reminiscence work, all of which have different definitions. Gibson (2006) takes issue with the term ‘reminiscence therapy’ stating that “reminiscence is not a therapy, which implies expertise and distance, but is better described as reminiscence work which, she claims, illustrates its participative nature” (Parker, 2006, p.45).

Similar to reminiscence, there is inconsistency around the terms for the prompts used to foster reminiscence, for example: ‘artefacts’ (Lazar et al., 2016), ‘memory triggers’ (Pöllänen & Hirsimäki, 2014), and ‘memory objects’ (Marschall, 2019). Multi-sensory prompts are important in the reminiscence process for promoting recall and enriching the experience. These prompts (which later we will coin ‘associative items’) stimulate a response and may include photographs, music, knitting patterns, smells, tastes, textures, and old objects (Kasl-Godley & Gatz, 2000; Lin et al., 2003). Gibson (1994) differentiates between ‘general’ and ‘specific’ reminiscence when selecting ‘memory triggers.’ In ‘general’ reminiscence work, a variety of theme-based multisensory triggers can be used. This takes into consideration the known background and interests of the participants, while in ‘specific’ reminiscence, the triggers are carefully selected to reflect the person's detailed life history (Gibson, 1994).

Both music therapy and reminiscence work are widely used in dementia care and were highlighted as valuable in promoting well-being in the NICE Guidelines updated quality standards on dementia (NICE, 2019). Music therapy enables the promotion and expression of personhood, while providing ample opportunities for enriched social, emotional, and therapeutic relationships. It is a process-orientated intervention that focuses on the present moment, matching the client’s level of awareness and cognition (Robertson-Gillam, 2011). Reminiscence work can promote an awareness of self-orientation to place and time, and a sense of reason and meaning in life (O’Rourke et al., 2011). The overlap between music therapy and reminiscence work is rather pertinent in that they are intrinsically person-centred and have the capacity to facilitate people with dementia to recall autobiographical memories tied to their self-identity and life story. Music therapy enhances the experience of reminiscence by incorporating the significant music in people’s lives, the rich associations that lie within, and creative musical expression.

Although ‘reminiscence music therapy’ is cited sparsely in the literature, and a small number of studies have demonstrated positive effect, previous studies lack rigour and consistency (Istvandity, 2017). Furthermore, there is a lacuna regarding facilitation and the use of ‘associative items.’ This naturally points towards a focused, combined intervention with the use of associative items, which we will refer to as Reminiscence-Focused Music Therapy (RFMT), and how it might promote positive mood and engagement levels, shared interaction, and reminiscence for people with dementia.

After observing the improved mood, engagement, and alertness that older people experienced when they attended a therapeutic choir on clinical placement at a residential care facility, the first author was curious to know how alternative methods could encourage these responses in people with dementia. The second author’s (research supervisor) involvement in a previous sensory intervention study (Ennis, 2013), which promoted increased alertness, positive mood, and engagement for people with dementia, was an influencing factor which informed the current study. A discussion was ignited between the authors on how group music therapy and reminiscence combined with ‘associative items’ might increase mood and engagement and promote shared interaction. Subsequently, our hypothesis is as follows: The combination of music therapy and reminiscence work with associative items will stimulate the senses with the intention of evoking emotions and triggering related memories.

The aims of this study are follows:

1. To explore if RFMT can temporarily promote positive mood and engagement levels and shared interaction through (a) musical expression, (b) reminiscence, and (c) verbal interaction.

2. To reflect and gain a deeper understanding of the intervention considering the music input, the role of the associative items, and experience of the facilitator.

3. To provide a comprehensive overview of the design and implementation of RFMT.

4. To explore the implications of practice and provide a facilitation guide for music therapists.

The study was reviewed and approved by the Research Ethics Committee (REC) in the Faculty of Arts, Humanities and Social Sciences at the University of Limerick, Ireland. Initial consent for the research was sought with information provided in an appropriate way to help the participants fully understand the study. It is important that people with dementia are supported and take part in research if they wish to do so (Higgins, 2013). Assent was also received by their next of kin. Maintaining consent can be an issue when working with people with cognitive impairment, due to memory changes, and it may need to be revisited and re-established (Dewing, 2007). “Provision must be made to repeat consent for each research encounter if necessary” (Dewing, 2002, p. 19). The participants have been represented by pseudonyms in this paper, and any identifying information has been excluded to ensure anonymity.

The inclusion criteria of this study required that participants have a diagnosis of moderate Dementia with a score of 13 > 20 in the Mini-Mental State Examination [MMSE] (Folstein et al., 1975). They also were required to score in either the ‘planned’ or ‘exploratory’ level of the Pool Activity Level Assessment [PAL] (Pool, 2012). The PAL is a strengths-based tool that assesses a person’s cognitive and functional level through everyday activity. These criteria were vital in ensuring that participants had the capacity to engage interpersonally with the researcher and other participants. The facilitator (Author 1) conducted an individual music therapy assessment session with each participant prior to the commencement of the study, to confirm cognitive and functional levels and their capacity to engage and communicate. It also explored the individual's biographical history, preferred music, and significant songs, and was used for rapport-building. The participants of this research study were 5 females in long-term residential care with a diagnosis of moderate dementia with an MMSE of 13 > 20 and a ‘planned’ or ‘exploratory’ PAL level.

Five themed sessions were planned for the RFMT programme. Themes and corresponding music and associative items were brainstormed and chosen beforehand. The themes included: Christmas, Spirituality, Autumn, Going to Dances, and Childhood (Table 1). Taking a person-centred approach, the chosen songs considered participants preferred music.

In each of the sessions the same greeting song was used to welcome and orient group members. Similarly, a familiar goodbye song was used to close the session each week. The structure of the session was as follows: (a) verbal conversation, (b) singing of a familiar song, (c) introduction of the associative items, (d) time for interaction with associative items, (e) repeat chorus of familiar song with associative items, and (f) move on to the next song. Four songs and four groups of associative items were chosen for each RFMT session to allow adequate time to interact with each participant and allow sufficient time for responses.

The RFMT programme took place between November 2019 and January 2020. It was an exploratory and procedural study with a focus on establishing the methodology; therefore, no control sessions were compared to the RFMT intervention. The first author acted as the facilitator and delivered the RFMT sessions, while the second author acted as an observer and ‘mapper’ and carried out the Dementia Care Mapping using DCM-8 guidelines (Bradford Dementia Group, 2005b).The observer had completed the approved 4-day DCM course, which included an assessment of competence in the use of the tool. He achieved over 70% interrater reliability with a DCM trained colleague, conducted over 50 hours of mapping in residential care, and facilitated mapping for a previous research study (Ennis, 2013).

DCM involved the mapper taking observations of a small group of individuals at 5-minute intervals. Behaviours of the group members (talking, self-care, leisure, expression, reminiscing etc.) were systematically coded from a possible list of 24 behaviour code categories (BCC). The behaviours were mutually exclusive; therefore, an individual could only be recorded engaging in one behaviour in any one particular time frame. If multiple behaviours occurred, operational rules prioritise behaviours with greater potential for higher mood and engagement. For example, ‘Reminiscence’ or ‘Expression’ takes precedence over passive engagement or being withdrawn. Simultaneously the mapper coded the mood and engagement value (ME). The ME value assessed the mood/affect and engagement experienced by each person during every 5-minute time frame. Mood was scored on a six-point scale from –5 (great signs of negative mood) to +5 (high positive mood), while engagement was scored from –1 (withdrawn) to +5 (deeply engaged). These scores were then tallied to arrive at the coded ME value. Although individual behaviours and ME values were recorded and reflected upon, they are not presented in this paper due to intermittent attendance.

Contemporaneous field notes taken by the second author recorded responses of the group members to stimuli (musical, visual, kinesthetic, tactile, olfactory, gustatory) and the associative objects. In addition, they captured musical expression, reminiscence, and verbal and non-verbal interactions that occurred between group members and their peers, and between group members and the facilitator. Post session, these notes were cross-referenced against the behavioural codes and ME scores, and further developed with the facilitator to provide comprehensive session summaries.

In the final step of the process, the facilitator completed a reflective journal for each session. This aimed to acknowledge the subjective nature of the researcher’s interaction and interpretation of the data, provide a decision-trail within the public domain, and transparency of the processes leading to conclusions being presented (Jasper, 2005).

The DCM data were imported into a specifically designed excel-based programme, which supports automated generation of individual and aggregate scores and graphs for reporting (Bradford Dementia Group, 2005b). Data from the session summaries was analysed with the intention of identifying moments of (a) musical expression, (b) reminiscence, and (c) verbal interaction. These session summaries were cross-referenced with the DCM coding and later colour coded and compared with the qualitative data.

Qualitative content analysis (Hsieh & Shannon, 2005) was used to analyse the data from the facilitator’s reflective journal. This involved a systematic classification process of coding and identifying themes or patterns from the text data. To avoid preconceived categories, the analytical process followed four main stages: reading data thoroughly, rereading and deriving code from the text, making notes of impressions, thoughts and initial analysis, and categorisation of codes based on relationships/links (Hsieh & Shannon, 2005). To avoid potential bias, the data were additionally reviewed by the second author for congruence between data, codes, and themes.

RFMT was highly effective in promoting positive mood and engagement levels (+3 = 51.6%, +5 = 38.2%), and no negative scores were recorded. Qualitative data from the reflective journal and session summary notes supported and illuminated the DCM findings. The mood of the group members, although remaining positive, was changeable throughout the sessions. Sometimes group members were more reflective and sombre, especially during the ‘Spirituality’ session when discussing difficult memories or the loss of a loved one. However, negative feelings were short-lived, as the music or associative items redirected the conversation. On other occasions, the mood was playful as group members interacted with associative items or danced in their chair to a familiar song. The wide range of moods recorded is reflective of the flexible nature of the RFMT sessions, which were client-led and person-centred.

Group members remained actively engaged in the RFMT sessions through singing, reminiscence, verbal conversation, and interaction with the associative items. The act of group singing kept the group members highly engaged and alert, as well as promoting a positive mood. On several occasions during reminiscence and verbal interaction, some group members became overstimulated. This led to periods of disorientation or disengagement with the activity. Song is known to be a modality of connection, even as we live with the changing constellations of cognitive impairment (Parmar & Puwar, 2019). In the RFMT session, the music acted as an anchor and successfully re-engaged and reoriented group members to the present moment.

When designing the sessions for this study, songs were chosen by the authors based on group members' preferred music. However, group members often spontaneously contributed songs from their past that were prompted by the theme of the session or the associative items. For some members of the group different songs evoked personal associations. Tomaino (2002) stated that people have memories not only of the details of a song (e.g., melody or lyrics), but also of the memories and rich associations related to the song. This occurred in the ‘Childhood’ themed session when one group member reminisced about how her childhood friend disliked the song ‘Molly Malone’:

Similarly, another group member spontaneously sang ‘The Power of Love’ during the ‘Childhood’ themed session, conforming to the well-known phenomenon of the ‘reminiscence-bump’ (Baird et al., 2018). The member in question had rich associations with the song and she recounted that she performed the song live on the radio when she was younger. Studies have demonstrated that music can trigger autobiographical memories with emotional content (Cady et al., 2008; Clements-Cortes & Bartel, 2018; Tomaino, 2002), and a person’s memories of music are closely linked to their identity and life story (DeNora, 2000; Istvandity, 2014a). Each individual has their own unique ‘songprint’ which is the repertoire of significant songs or music we each carry, influenced by our culture, age, and personality (Vander, 1996). This was evident in this study as group members spoke of memories associated with songs from their earlier days.

The act of group singing provided a sense of togetherness and group cohesion. In contrast to turn-taking social activities such as talking or reminiscing, music can be facilitated with everyone simultaneously expressing themselves, while every individual is still heard (Brancatisano & Thompson, 2020). When working with people with dementia, singing is one of the most powerful ways to engage and build a connection with others (Tamplin & Clark, 2020). Music can bring people with dementia back into the social space and enables them to be responsive to others, to meaningfully participate, and “to restore their status as socially recognizable actors” (Matthews, 2015, p. 575). The music acted as a supportive framework holding the therapeutic space, while the act of group singing was mindful and connected the members through a positive social experience and shared memories.

A multitude of associative items were used throughout this study, incorporating auditory, visual, tactile, olfactory, and gustatory stimuli to elicit memories and encourage musical expression, reminiscence, and verbal interaction. As no standardised protocol for choosing associative items existed, the authors brainstormed and consulted with other colleagues in relation to the selection of associative items. This exploratory study informed the authors of the responses that certain items elicited and their appropriateness in this context. Certain associative items had sensory appeal, such as the fragrance of perfume, the texture of sand and seashells, and touching pearls and fabrics. However, ultimately the function of the associative items was not to directly engage the senses but to elicit memory and encourage shared interaction. This perhaps suggests why sensory stimulation (T) was only recorded for 2.6% of the overall time during this study while reminiscence was recorded for 23%.

While the associative items predominately encouraged reminiscence, they fostered verbal interaction and some members used detailed language to describe them. They also stimulated cognition, and participants attempted to recall childhood game rules, recipes, and the structure of the rosary. These can be referred to as procedural memory, the ability to unconsciously recall information we learn by rote, such as tying one’s shoe, playing an instrument, singing a song, and praying and are often enduring skills in people with dementia (Klimkowicz-Mrowiec et al., 2008; Poldrack & Gabrieli, 1997).

Furthermore, some of the associative items [photographs] supported reality orientation and acknowledgement of present circumstances and feelings:

The incorporation of overarching themes in this intervention aimed to provide a platform for musical expression, reminiscence and verbal interaction. Although reminiscence often occurred organically upon presentation of the session theme, different memories were elicited for each group member, some of which were related to a song or an associative item.

RFMT allowed group members to socially interact and share stories from their past, which were stimulated by the chosen music and associative items. The supportive social environment of the music therapy space enabled group members to communicate both musically, gesturally and verbally, maintain social relationships with their peers, and be included in a group activity despite their dementia. The small group size was invaluable in allowing members to mutually support and comfort one another while reminiscing (Bright, 1997a, 1997b). RFMT encouraged group members to engage in reminiscence with their peers and the music therapist. The arts are particularly appropriate for people with dementia because they give a lot of scope for non-verbal communication, emotional expression, opportunities for creativity, and to sit back and enjoy the humour and originality of other participants (Schweitzer, 2011).

Spirituality was experienced in the sessions, through musical expression, prompted by the associative items and supported by reminiscence. Experiencing spirituality can give people the ability to look for meaning and purpose in life, instill hope and growth, and provide an inner force to go beyond one’s personal circumstances (Parker, 2007; Speck, 2005). Furthermore, it has been shown to increase well-being and reduce loneliness and low mood (Toivonen et al., 2017). Spirituality is integral to working with people with dementia and has increasingly been explored in the music therapy literature in recent years (Kirkland et al., 2014; Magill, 2002, 2006; Moss, 2019; Potvin & Argue, 2014; Tsiris, 2016). Recurring spiritual themes that have emerged include relationship, remembrance, prayer, and peace (Magill, 2002, 2006). For people with dementia, spiritual care needs can often go unmet due to cognitive and communication deficits; however, many people with dementia can demonstrate their faiths verbally, physically, and emotionally (Mooney, 2004). Music is another key way that faith can be expressed and experienced. This was strikingly obvious in the ‘Spirituality’ session where there was increased musical expression as group members independently initiated several Irish Catholic hymns such as ‘The Bells of the Angelus’, ‘Queen of Heaven’, ‘Holy God’, and ‘Hail Queen of Heaven’. The most prominent vocal strength was noted during these musical interactions and provided a sense of togetherness, allowing group members to interact with and support one another.

These group members grew up in the heart of the traditional Catholic faith, convent schools choirs, Corpus Christi, May processions and Sunday mass. While singing these remembered and implicitly known hymns, frequent eye contact within the group was observed. Furthermore, it appeared to evoke a sense of self confidence and identification with one another, the familiarity of their shared faith providing a sense of security and comfort.

Some of the music and associative items had powerful connotations for members of the group. For example, when holy water was offered on their finger, all members automatically blessed themselves, making the sign of the cross. On two occasions unprompted religious responses were observed: L looked at the miraculous medal and said, “Oh Mary conceived without sin, pray for us.” and S touched the rosary beads and said, “Oh Mary of Fatima, pray for us.” These instinctual responses were automatic, ingrained from a lifetime of practice, and led to unprompted initiation of prayer and the expression of their faith. There was a reverence for the items that led to a more contemplative experience and private communication with oneself.

Topics of spirituality came up across many of the sessions related to Catholic schooling and religious ceremonies. For many older people in Ireland, the Catholic Church was pivotal in their upbringing and continues to be central in their lives. Therefore, spiritual care is as important as physical and psychological interventions (McSherry, 2006), and spiritual needs are core elements of comprehensive healthcare and music therapy practice (Moss, 2019).

The process of implementation and facilitation was key to the effective delivery of RFMT. Kitwood (1997) emphasized that the aim of facilitation is “to enable interaction to get started, to amplify it and to help the person gradually to fill it out with meaning. When this is done well there is a great sensitivity to the possible meanings in a person’s movements, and interaction proceeds at a speed that is slow enough to allow meaning to develop” (p. 109). When designing and implementing this intervention, the emphasis was placed on the structure, content, and delivery of the music and associative items, as well as allowing adequate time for reminiscence and verbal interaction. Adaptability was an essential skill required on behalf of the facilitator when certain music or associative items did not elicit a response or if the topic of conversation was not related to the session theme. The appropriate pacing and allowance of space was imperative. It supported individual cognitive levels and afforded opportunities for initiation and creativity.

The successful facilitation of reminiscence work within a music therapy framework involves blending best practice in person-centred music therapy and experience in facilitating reminiscence. Due to the social and psychological consequences of dementia, we must work in ways which enable people with dementia to participate without fear of failure, by recognising their difficulties and navigating ways around them. It is essential to play to their remaining strengths while being sensitive of their feelings (Schweitzer & Bruce, 2008). This aptly summarises some of the implications of practicing RFMT and highlights that facilitation is key in making this approach successful. It must be noted that difficult topics were discussed such as death, loss of independence, and negative childhood experiences. Although these discussions were short-lived and no negative mood or engagement levels were recorded, it was essential that the facilitator was equipped to manage these conversations and hold a safe space for the other group members. In developing and evaluating reminiscence techniques it is vital to understand or predict how revisiting the past might affect people (Bluck & Alea, 2002).

RFMT may also be used successfully with people with more advanced dementia. “Recognition persists long beyond the point, where people can remember and relay their memories independently,” and therapists can foster recognition with the sensitive use of individual biographies alongside specific associative items (Schweitzer & Bruce, 2008, p. 26). People respond with pleasure to their own life story or familiar accounts given by others and this may lead to the spontaneous contribution of some other memories. Similarly, RFMT may also be used with individual clients rather than in a group context to work in a deeper or more focused way and promote a sense of identity through musical expression and reminiscence. In this case, associative items may be personal to the individual and hold more meaning.

Due to time constraints, it was beyond the scope of this study to compare RFMT to traditional approaches of both music therapy and reminiscence work or to standard care alone. Future research may compare the results of RFMT with other controlled measures. This exploratory study also had a small sample size and therefore the findings should be interpreted with care. Future studies could compare behaviours observed during the intervention with pre-intervention behaviours, while mapping behaviours post intervention could evaluate persistence effect.

DCM is noted to be “suited to smaller scale within-subject or group comparison intervention evaluations”; however, due to its coding rules, it has been criticised for a bias toward highlighting positive behaviours and underestimating socially passive and withdrawn behaviours, unlike data collected with continuous time sampling, which would capture all behaviours equally (Bradford Dementia Group, 2005a, p. 115; Kirkland et al., 2014). DCM prioritises and highlights positive behaviours, because its aim is to identify potential, to build upon good practice, and encourage person-centred care.

It would be of interest to investigate spontaneous dyadic interactions with an RFMT group when compared to standard care alone. In a more controlled study with a strictly defined protocol, spontaneous verbal interaction within the group could also be examined in more detail, to measure the effect of the facilitator.

The first author had a complex dual role as both a researcher and facilitator. The research involved reflecting upon personal experiences of facilitating the sessions, and theorising and contextualising this experience. If this study were quantified in the future, the authors would recommend that the data be collected and analysed independently of one another by separate researchers to avoid subjectivity and bias.