Participants were a mother (Natalia)¹ and her preterm daughter (Ana), who had no syndromes or congenital anomalies and was admitted to the NICU of a public hospital in Porto Alegre (State of Rio Grande do Sul, Brazil).

Natalia was 24 years old; she lived in Porto Alegre, had concluded primary education, and was unemployed. She had three children: two of them, 1 and 3 years old, were from previous relationships, whilst Ana was the daughter of her third partner who had been incarcerated almost a year earlier. The family had a low socioeconomic income. Ana was born at 27 weeks² of gestational age with a weight of 685 g. The Apgar Score³ at birth was 2 (1st minute), 4 (5th minute), and 6 (10th minute). After birth, due to symptoms of sepsis and respiratory distress, Ana was put in an incubator and submitted to orotracheal intubation. Besides the preterm birth, the infant presented no other severe clinical condition. After 118 days of hospitalization in the NICU, Ana had a weight of 2620 g at discharge; she could breathe autonomously and breastfeed.

A single case study design (Stake, 2006) was used, with five phases of data collection. In Phase 1 (Pre-intervention), 20 days after the infant's birth, the mother was invited to participate in the study, she signed the Informed Consent Form and took part in the following interviews: the Interview about motherhood in the context of prematurity (NUDIF/PREPAR, 2009a), which was used to investigate the experience of motherhood in the post-partum period; and the Interview about the mother's musical history (Palazzi, Meschini, & Piccinini, 2014a), which investigated the acoustic environment and the mother's experiences and musical preferences throughout life and during pregnancy. Both interviews were structured but conducted in a semi-directed manner. Moreover, a Demographic data sheet (NUDIF/PREPAR, 2009b) was filled out in this phase together with a Clinical data sheet for the mother and her preterm infant/post-partum (NUDIF/PREPAR, 2009c), which was updated at each phase of data collection.

In Phase 2 (Intervention), a week after Phase 1, the mother participated in the Music Therapy Intervention for the Mother-Preterm Infant Dyad – MUSIP (Palazzi, Meschini, & Piccinini, 2014b). The MUSIP is a music therapy intervention that aims at sensitizing the mother to the importance of singing to her infant and supporting the mother-infant dyad during maternal singing. It was created for this study based on music therapy research on prematurity. The MUSIP is organized in eight sessions, divided between meetings with the mother and sessions with the mother-infant dyad in the NICU. However, the structure of the MUSIP had to be adapted in this study in order to meet the mother's needs and the space availability in the NICU as described below. Besides this, due to Ana's clinical instability, the sessions with her started at 31 weeks of post-menstrual age⁴.

Session one only involved the mother, was conducted in a room of the Neonatology Unit and lasted approximately 45 minutes; sessions two, three, four, and five were carried out in the NICU with the mother and the infant in the incubator; furthermore, sessions six, seven, eight, and nine were held with Natalia during skin-to-skin contact with Ana (kangaroo care). The sessions with the dyad in the NICU lasted 15 to 20 minutes. In session one, the mother and the music therapist sang some songs selected by the mother (“Fico assim sem você” [“I stay like this without you”] by Claudinho and Buchecha and the Brazilian lullaby “Nana nenê” [“Rockabye baby”]), with and without guitar accompaniment. Sessions two, three and four aimed at accompanying and supporting the mother during the experience of singing for the daughter, by stimulating the observation of the infant's responses and by emphasizing the interactive potential of singing. Sessions five and six focused on the improvised humming with the dyad in the NICU and lastly, sessions seven, eight, and nine aimed at encouraging the mother to write a song for the infant.⁵ A music therapist, the first author of this study, conducted the nine sessions of the MUSIP (approximately once a week for 2 months) with the clinical supervision of another music therapist, the second author of this study.

In Phase 3 (Post-intervention), a week after the end of the MUSIP, the mother took part in the Assessment interview of the music therapy intervention for the mother and her preterm infant (NUDIF/PREPAR, 2014b), which investigated the impact of the intervention on maternal feelings about the music therapy, the therapist and her satisfaction with the intervention, as well as her perceptions about the changes in the infant and in their bonding. It was a structured interview conducted in a semi-directed manner.

Furthermore, the dyad participated in a session of the Observation of the mother-preterm infant interaction (Palazzi & Piccinini, 2014), during maternal singing in the NICU. This session aimed at observing infant's and mother's behaviors and their interactions, during singing and non-singing episodes. The mother was oriented to interact with the infant approximately 2 to 8 minutes during each episode. The music therapist, with the help of a psychology student, videotaped the observation sessions by using a GoPro Hero 3+ camera within the NICU, and a Sony DCR-sr85 camera out of the NICU. The video footage focused on mother-infant interaction.

In Phase 4 (Pre-discharge), carried out before the infant's discharge, the mother took part in another assessment interview adapted for the pre-discharge phase, and a new observation session was conducted during breastfeeding in the NICU. To conclude, in Phase 5 (Post-discharge), a week after discharge, a new observation session was held in a hospital room in two different moments: during maternal singing and during non-singing while changing the diaper. The first author carried out all the MUSIP sessions, the observation sessions in Phase 3 and 5, and all maternal interviews in Phase 1. A psychology student conducted the observation session in Phase 4, as well as all the clinical data sheets of the infant throughout the phases, and conducted the assessment interviews with the mother in Phase 3 and 4. All the observation sessions and the MUSIP sessions (with the exception of session two) were audio or videotaped. Two psychology students were responsible for transcribing maternal interviews. This study was approved by the Ethics Committees of the Psychology Institute of the Federal University of Rio Grande do Sul (UFRGS) (n. 985.941) and of the hospital (n. 1.069.283).

To analyze data, we examined the mother’s interviews in the pre-MUSIP (Phase 1), post-MUSIP (Phase 3) and in the pre-discharge (Phase 4) period, as well as the video footage of the mother-infant dyad during MUSIP (Phase 2) and during the observation sessions (Phase 3, 4, and 5). Initial interviews were used to describe the case. The assessment interviews in the post-MUSIP and in the pre-discharge period were examined through thematic analysis (Braun & Clarke, 2006), with the aim of investigating maternal perceptions about the music therapy contributions to the mother-preterm infant dyad in the NICU, in particular to the infant, the mother and the mother-infant interaction.

Video footage of MUSIP sessions were analyzed as follows: 1. First, we watched the video footage and wrote a general report on all sessions; 2. Secondly, we selected three of the six videotaped sessions (sessions four, six, and nine)⁶, and we identified and selected the first non-singing episode (3-4 minutes) and the last maternal singing episode (3-8 minutes) of each session, with a total of approximately 14 minutes of singing and 11 minutes of non-singing⁷; 3. Lastly, the first author wrote detailed descriptions regarding infant's and mother's behaviors and mother-infant interactions during singing and non-singing episodes. The initial and final time of each behavior under observation were noted, so as to also have recorded information about the duration of infant's and mother's behaviors and their interactions.

We used a similar procedure for the video footage concerning the sessions of the Observation of the mother-preterm infant interaction (Phase 3, 4, and 5). 1. First, we identified and selected the last 3 minutes of each video excerpt, with a total of approximately 6 minutes of singing and 4 minutes of non-singing⁸; 2. Secondly, the first author wrote detailed descriptions regarding infant's and mother's behaviors and mother-infant interactions during singing and non-singing episodes, highlighting the initial and final time of each behavior under observation, so as to have information about the duration of infant's and mother's behaviors and their interactions. A psychology student subsequently reviewed all video excerpt descriptions in an attempt to ensure consensus. Interviews and video excerpts contributed to data triangulation (Stake, 2006) allowing for an in-depth analysis of the case, based on maternal perceptions, on infant's and mother's behaviors, and on their interactions.

Data were integrated and analyzed through thematic analysis (Braun & Clarke, 2006), using a deductive approach based on the structure of themes and on the theoretical background of Haslbeck (2013, 2014). The analysis was based on two themes: a) empowerment of the infant and b) empowerment of the mother. In particular, the interviews and the descriptions were read several times, so as to identify vignettes and excerpts of the descriptions that portrayed the themes used for analysis.

Haslbeck’s (2013, 2014) structure of themes was originally created from studies that involved the therapeutic process between the music therapist and the infant. Therefore, this structure was adapted for the purposes of this study in order to consider the interactional process between mother and preterm infant. The original themes that referred to the music therapist's behaviors were adapted and employed to analyze maternal behaviors. The original structure also involved other themes, such as the responsiveness between the music therapist and the infant, which was not used in this study, since this category and all the dimensions about mother-infant interactions were included in the themes of empowerment of the infant and empowerment of the mother. Furthermore, the original structure involved the theme of communicative musicality, which was also investigated in this study, but it will be presented in a future publication. In conclusion, we added other characteristics to the themes based on the literature being specifically derived from data from this study.

According to Haslbeck's definition (2013, 2014), empowerment of the infant refers to the infant's ability to relax, calm down, and enhance self-regulation, orientation, and interaction. This theme involves the infant's relaxation, the presentation of new competences, the participation and the engagement in singing, and for the purposes of this study, we also included the stabilization of oxygen saturation.

In the first MUSIP sessions, the infant was still unstable, showing frequent agitation and saturation drops. Therefore, in session four the empowerment of the infant could not be identified during singing and non-singing. However, in session six, Ana was more relaxed and in quiet sleep state during both singing and non-singing. Moreover, in session nine, singing contributed to a greater relaxation of the infant when compared to the non-singing episode. In this session the infant showed new competences, as well as participation and engagement during singing. In fact, while her mother was singing, the infant was in an alert quiet state, moving herself and vocalizing:

In the post-MUSIP and pre-discharge period, the empowerment of the infant appeared more evident. For example, during the assessment interview of the intervention, Natalia reported that the MUSIP sessions and her spontaneous experiences of singing helped her daughter calm down when she was agitated,

Besides enhancing the infant's relaxation, the MUSIP contributed to stabilizing Ana's oxygen saturation since the mother reported that when she sang, the saturation increased and kept stable. Natalia emphasized the joint positive effect of singing and kangaroo care on re-stabilizing oxygen saturation: “When her saturation starts dropping, I ask to pick her up, I sing and I rock her, and then she stabilizes” (Int3). Moreover, the mother reported that when she talked to Ana, her saturation kept decreasing, while when she sang the saturation remained more stable:

Relaxation during maternal singing was also reported in the descriptions of the observation sessions in the post-MUSIP. In fact, the infant slept throughout the session, without displaying any signs of agitation, while the mother hummed improvised tunes, rocked, and caressed her. The mother also reported that after the music therapy sessions, the infant seemed to recognize the maternal voice and react to singing, by opening her eyes and smiling:

In the pre- and post-discharge period, both the interview and the descriptions of the observation sessions highlighted the empowerment of the infant through the presentation of new competences and the engagement in singing, more than through relaxation as it was shown in the post-MUSIP. In fact, during singing and non-singing the infant interacted actively through her gaze, prolonged face-to-face contact with her mother, and vocalizations:

During singing there were more prolonged mother-infant gazes and also wider movements of the arms, diversified vocalizations, and several facial expressions. The infant looked at the mother more frequently when she was singing compared to the non-singing episodes:

According to Haslbeck's definition (2013, 2014), the empowerment of the mother refers to her ability to relax, calm down, interact with the infant, and increase maternal competences. This theme involves mother's relaxation, the strengthening of maternal competences, and in this study, it also includes the overcoming of the embarrassment and the fear of interacting with the infant and the autonomy in singing.

During the first MUSIP sessions, the mother was anxious and agitated when she sang to her daughter. However, in session four she prepared herself for the singing experience by closing her eyes and leaning on the incubator, which was identified as an attempt to relax. The descriptions of session four showed the mother's autonomy in singing, since she hummed for her daughter without waiting for the music therapist:

The mother's relaxation was more evident in the descriptions of singing episodes related to sessions six and nine. While she was singing during kangaroo care with the infant, Natalia closed her eyes and leaned her head or mouth on Ana’s head for a longer period. This behavior was not present in the non-singing episodes of the MUSIP, suggesting that Natalia probably associated this position with the experience of singing to her daughter. Furthermore, in these sessions, she overcame the embarrassment and the fear of interacting with her daughter, which have been reported as the mother's initial difficulties in the first interviews. She also showed maternal competences, by touching her daughter, caressing her, and being responsive to the infant's behaviors:

In the non-singing episodes of the same sessions (six and nine), the strengthening of maternal competences was also perceived. The mother touched her infant and used expressions of affection, but she altered her gaze focus more frequently. Sometimes, she looked at her daughter, while other times she looked at the NICU:

Lastly, we observed greater autonomy of the mother in singing, compared to the prior sessions, since Natalia hummed improvised tunes to her daughter:

In the post-MUSIP, the empowerment of the mother through relaxation appeared both in the interview and in the observation session. In fact, the mother reported that she “felt good” and more “relieved” during music therapy and when she sang to Ana. She highlighted that music therapy helped her to “loosen up,” “feel less embarrassed” of talking to her daughter, and overcome the fear of touching her, since singing was always associated to touching or kangaroo care. This contributed to “build a bonding” between the mother and the infant, to stimulate Natalia's interest for her daughter and to help her “taking care” of Ana:

Furthermore, music therapy contributed to the empowerment of maternal competences and the mother's autonomy in singing, since Natalia integrated the improvised humming in her daily routine at the NICU, by using singing as a resource to calm her daughter down when she was agitated or when the oxygen saturation dropped:

Similarly, the descriptions of the observation sessions showed that Natalia sang autonomously to her daughter, with improvised humming, and adding many prolonged and diversified expressions of affection:

In the pre-discharge interview (Phase 4), the empowerment of the mother was more visible through the strengthening of her competences and her autonomy in singing, than through relaxation. In fact, in this phase the interactive potential of singing was more evident. Whilst answering a question made by the interviewer about the contributions of the music therapy, Natalia reported:

Despite this, in the pre-discharge period Natalia reported that she did not sing anymore for her daughter, because Ana was no longer agitated, she could pick her up more often, thus preferring to just talk to her.

In fact, in the pre- and post-discharge period, in the non-singing episodes, the mother talked to her daughter through infant-directed speech¹¹ that is contingent with the infant's behaviors, such as gaze and sucking:

Even though in the pre-discharge period the mother stopped singing for her daughter, she reported that she wanted to sing at home after discharge with the whole family when the infant would be agitated to calm her down and help her to sleep. According to Natalia, music therapy seemed to represent a transitional stage that helped her to awaken her maternal competences, to connect with the infant, and talk more to her. In the post-discharge (Phase 5) period, we observed that when Natalia sang she dedicated more exclusive attention to her daughter, showing more diversified expressions of affection and allowing Ana to more face-to-face contact and interactions:

In the pre- and post-discharge period we observed that the mother was responsive to her daughter's behaviors, for example imitating her facial expressions, answering to the infant’s visual contact and stopping singing to wait for her daughter's answers:

(00:07) The infant makes visual contact with the mother and the mother says “Shush” to her. The infant looks at her.

Nevertheless, during non-singing episodes the mother alternated her gaze focus, sometimes looking at her daughter and other times at the NICU environment, as previously pointed out:

On the contrary, when the mother sang, the expressions of affection and face-to-face contact were more prolonged. In fact, after discharge, the mother looked at the infant during the whole observation session:

Before concluding, it is worth mentioning some limitations of this study. First, it was impossible to fully meet all the initial eligibility criteria, due to the complexity of the clinical conditions associated with prematurity. For example, the mother was not primipara and was already used to singing to her children. This might have contributed to strengthen the benefits of the intervention. On the other hand, she did not live with the father of the infant, had a limited availability of time, was worried about her children at home, and had little family support. All these factors might have made her engagement in the intervention more difficult. In addition, we have evidence from just one dyad, so it is important to adopt the MUSIP with a bigger sample.

Secondly, we could not investigate maternal anxiety and depression in the pre- and post- intervention, and we suggest to examine it in future studies. The presence of maternal depression might have complicated the intervention, by limiting the mother’s engagement in singing to her daughter and in the interaction with her.

Thirdly, we adapted the application of the MUSIP, due to several conditions involving the infant, the mother, and the NICU, and we carried out just one individual session with the mother and all the other sessions with the mother-infant dyad. Despite being a limitation, the adaptation of the MUSIP also represented an advantage that made a more individualized intervention possible, suited to the needs of the mother-infant dyad. This might have contributed to strengthen the role of the mother in the interaction with her daughter. In fact, it is important to carry out interventions that are close to the clinical reality (Ettenberger et al., 2014) and, therefore, the application of the MUSIP requires the music therapist's sensitivity to the needs of the infant, the mother, and the NICU. Furthermore, it is important to highlight that this study was carried out in a public hospital in Porto Alegre, where the majority of the families of the hospitalized infants have a low income. Families in this socioeconomic situation may have more difficulties in getting involved in interventions. For instance, they may have problems in coming everyday to the hospital, because of the distance and costs of the travel. Mothers may have other children at home and may not have adequate social support for them while at the hospital. All this factors, together with personal difficulties to deal with the infant’s premature birth, might have affected the implementation of the original protocol of the MUSIP.

Fourthly, the structure of the themes used in the analysis (Haslbeck, 2013, 2014) was originally based on the music therapist-infant interaction and was adapted for this study to analyze the mother-infant interaction. The fact that the analysis was based on a deductive approach from two of Haslbeck’s themes might have contributed to disregard some aspects of our findings. Nevertheless, we tried to limit this risk, by adding to the themes some specific aspects emerging from the data of our study.

A qualitative approach was adopted to assess these themes, which allowed us to conduct an in-depth analysis, but it did not provide statistical data about the frequency of the observed behaviors. Lastly, the single case study does not provide an exhaustive comprehension of the phenomenon and the evolution of the mother-infant dyad was also due to factors such as the infant's development.