Comparison of a standard computer‐assisted cognitive training program to a music enhanced program: A mixed methods study

Abstract Background Between 17 and 75% of breast cancer survivors (BCS) experience long‐term cognitive deficits such as deficits in memory, attention, processing speed, and executive function. Aims This study aimed to (a) compare effects of a standard computer‐assisted cognitive training (CACT) program to a CACT program enhanced with music (CACT+A) to improve focus and concentration on BCS' memory, cognition, quality of life (QOL), and participation in everyday activities; and (b) garner participants' perspectives of effects of the programs to determine best practice. Methods An embedded design was employed in this mixed methods study. Participants who reported cognitive problems were recruited through breast cancer support groups. Four pre and post‐tests were used followed by a qualitative interview. Results Twenty‐five BCS, ages 31 to 72 years participated. The CACT group demonstrated significantly improved pre to post‐test scores for working memory, QOL, and three subscales of the Model of Human Occupation Screening Tool (MOHOST) measuring participation in everyday activities. The CACT+A group had significant improvement for four FACT‐Cog cognitive function subscales and the total score. Five themes emerged from the interview: Cognitive skill, Strategy learned, No change, QOL factors, and Participation in everyday activities. The CACT+A group expressed experiencing a larger ratio of improvements, most notably for memory and QOL factors. Conclusion CACT+A is an auspicious intervention option for BCS who self‐report cognitive issues. It is convenient to participate in at home and allows BCS to safely self‐isolate if need be. This study is a registered clinical trial protocol: TexasWU record 19 959.


| INTRODUCTION
Cancer is a continuing disease in our society with greater than 1.8 million new cases projected to be diagnosed in 2020. 1 In women, about 30% of newly diagnosed cancers will be breast cancers. 2 As of January 1, 2019, there were 3.8 million women breast cancer survivors (BCS) in the United States. 3 Researchers have estimated that between 17 and 75% of them experience long-term cognitive deficits such as deficits in memory, attention, processing speed, and executive function. 4 Such cognitive deficits of BCS are under diagnosed and can result in decreased participation in everyday activities. 5,6 Researchers 7-10 have shown rehabilitation can improve participation in everyday activities and quality of life (QOL). Rehabilitation generally consists of patients learning compensation skills to replace cognitive skills and/or remediation of cognitive deficits through practice of cognitive skills. Cognitive rehabilitation for remediation of cognitive skills can be provided to BCS with computer-assisted training. 8,[10][11][12][13] Previous computer-assisted cognitive training (CACT) programs for BCS focused on improving a variety of cognitive skills such as processing speed, 10 and executive function including working memory, cognitive flexibility, multitasking, planning, and attention. 8 Kesler et al 11 also concentrated on improving executive function and their intervention, consisting of computer exercises using visual stimuli, resulted in significant improvements in cognitive flexibility, verbal fluency, and processing speed, as well as marginal improvement in verbal memory. They conjectured that larger effects may have resulted in their study if a combination of visual and auditory exercises were employed, but auditory exercises were not available at that time. In 2019, the first author and a co-researcher compared a 4-week CACT program using primarily visual exercise to one with predominately auditory exercises. Both groups got 10 computer exercises for each exercise session but the primarily visual group was provided more of the visual exercises such as visual attention, visual and spatial memory, and visual memory, and the audio group got more auditory working memory, verbal memory, and verbal and visual memory exercises.
The researchers found that CACT with primarily auditory exercises did not result in any greater outcome scores than CACT with primarily vision exercises. Both groups demonstrated improved outcomes for perceived cognitive function and QOL for BCS. There was no control group. 12 The first author and a co-researcher have previously established that CACT improves outcomes for perceived cognitive function and QOL for BCS. 12 Remediation for cognitive deficits is dependent upon the premise that brain neuroplasticity can be achieved 14 ; computer brain-training exercises can facilitate plasticity. 15 In addition, some types of music can elicit specific brain waves that promote increased focus and concentration. 16 It is also likely that music interventions can affect brain plasticity due to its shared neural systems for reward, arousal, and affect regulation. 17 For this study, we were interested in investigating the effects on memory, cognition, QOL, and participation in everyday activities of BCS completing CACT enhanced with music to increase focus and concentration. We contended that in addition to using CACT to improve cognition and QOL in BCS, audio input of music designed to improve focus and concentration tasks should be examined, and if indicated included as a standard of care. The purpose of this study was 2-fold.
First we wanted to compare the effects of a standard CACT program to one with the same computer exercises, but which was enhanced with music to improve focus and concentration (CACT+A) on memory, cognition, QOL, and participation in everyday activities. Second, we wanted to garner participants' perspectives of effects of the two CACT programs to aid in determining best practice when using CACT. We hypothesized that a CACT+A would have greater change scores than a CACT on memory, cognition, QOL, and participation in everyday activities. We sought to answer the research question: "What are the effects of a standard computer-assisted cognitive training program compared to a computer-assisted cognitive training program enhanced with music on memory, cognition, quality of life, and participation in everyday activities perceived by breast cancer survivors?" 2 | METHODS

| Research design
A mixed methods study was employed in this study based on the belief that one data set would not be sufficient. 18 Utilizing this mixed methods allowed authors to combine elements of quantitative and qualitative methods to reach a greater depth of understanding and achieve corroboration of findings. 19 An embedded design was used with a qualitative strand embedded in the quantitative experiment to provide a secondary role of gaining participant perspectives. 18 Pre and post-tests were collected with outcome instruments and then a qualitative interview was performed. The major component was the quantitative data with a qualitative strand embedded after post-test quantitative data collection to allow understanding of how participants viewed changes and their experiences. Their feedback would be used to improve future interventions.

| Participants
A convenience sampling was used to collect the sample. Specifically, participants were recruited from three different breast cancer support groups and through one Facebook group that provided information to BCS. Once the participants were included in the study, they were randomly assigned into one of two groups. A priori power analysis was conducted using G*Power 3.1. With a desired level of power set at 0.80, an alpha (α) level at 0.05, and a moderate effect size of 0.30 (f) for repeated measures ANOVA (2 time points × 2 groups), it was determined that a minimum of 24 participants were required to ensure adequate power. The first author visited two of the support groups to present the proposed study and collect contact information of potential participants. A written description of the proposed study was sent to the director of the third support group and to the manager of the Facebook group. Potential participants contacted the first author and were given further information about the study. If they wanted to participate, a time and place for pretesting was arranged.
Inclusion criteria for participants were: must be a BCS and self-report cognitive problems, which they attributed to their breast cancer treatment. Exclusion criteria were persons who could not read or understand spoken English or who self-reported they had disorders that may affect their cognition including major mental disorder, central nervous system disorders, Alzheimer's disease, dementia, developmental delay, traumatic brain injury, or cerebral accident. Quantitative and qualitative data were collected from all study participants in both intervention groups.

| Instruments
Four quantitative and one qualitative outcome measures were employed. To measure working memory, the forward digit span was used. It has been shown to discriminate between BCS and controls. 20 It takes approximately 10 minutes to administer and has high reliability (0.891) for forward span and low for backwards span (0.598). 21  or restrict an individual's participation in everyday activities. We sought to determine whether strategies used to perform CACT would transfer to everyday activities. The MOHOST has shown good construct validity, item separation reliability, and concurrent validity. 24 We modified the MOHOST for our purposes by omitting the subset of communication and interaction skills as our intervention did not target these skills. Further, we considered the time to perform the pre  Table 1). Thereby participants were asked to note any changes in their memory, cognition, QOL, and in participation in everyday activities, they attributed to participating in the study.

| Research protocol
The study and consent form were approved through the authors' affiliated university institutional review board conforming to recognized standards of Declaration of Helsinki. Prior to any data collection, participants provided written informed consent at one of two sites based on participants' county of residence. Participants were randomly assigned to either the CACT or the CACT+A group, and administered the quantitative instruments described above which took approximately 45 minutes. Participants were then instructed on how to access and perform the computer-assisted exercises on the internet. Participants in the CACT+A group were issued headphones and an USB with a publicly available 2 hours and 30 minutes album of music championed to increase focus and concentration. All computer-assisted training was performed in a location of choice by the participant. Tablets were available if any participants did not have a device to access the Internet, but no one required one. An email was sent to the participants the night before their exercises began with their assigned password and a reminder to do their computer exercises for 30 minutes a day, 5 out of 7 days a week for 1 month. The total time of required computer exercises was 20 hours.
Computer software used for the CACT was from HAPPYneuronPro 25 which has nine different types of cognitive exercises including auditory, verbal and visual memory, verbal memory, executive functioning, processing speed, spatial memory, visual attention, visual memory, and visual and spatial abilities. Each exercise session was comprised of 10 different exercises run for 3 minutes. When a participant achieved a 100% percent score at a specific level of an exercise, they were given exercises from the next higher level of difficulty.
Participants' training was closely monitored to ensure that they completed the required amount of exercises. If they were short training time, they were sent a reminder email. Each week the number of exercises delivered from each type of exercise was modified based upon which cognitive exercise scores most needed improvement. The last week of exercises was selected based upon whatever types of exercises the participant had the lowest scores.
At the end of the month, all pretests were repeated, and the qualitative interview performed in person. The interviews were recorded on a digit recorder and notes about the participants' answers were written on an interview template. Participants who finished the study received $150.00 worth of gift cards as reimbursement. ing was applied to inductively identify concepts and themes, 26 and link participants' words to codes frequently using in vivo coding.
Constant comparisons were used next in axial coding to group concepts into categories, and theoretical questioning was employed to establish relationships between categories and subcategories. 26 Selective coding did not occur as it was not the purpose of this article to construct a theory.
Several validation and reliability strategies were used to increase trustworthiness. Creswell and Poth 27 recommend that at least two means of validation be used. Four validation strategies were used in this study. The first author revealed her bias in noting her previous research. She also had prolonged engagement with the participants and engaged in peer debriefing of the data and research process with a former co-researcher who is familiar with CACT and the population, BCS. Negative case analysis was demonstrated by indicating when some participants indicated that they did not feel any changes from participation in the study. Reliability of data was enhanced by transcribing the digit recording and reading transcripts for accuracy. 28 Transcripts were both coded line-by-line by hand and then imported into ATLAS.ti for more detailed coding.

| RESULTS
A total of 25 participants (

| Cognitive function
The FACT-Cog cognitive function subscales include perceived cognitive impairment (PCI), comments from others, PCA, and impact of cognition on QOL. As shown in Table 2, the subscale scores for PCI were the subscale scores were elevated significantly from pre-to post-test in both groups (ps < .05), but no significant difference was observed between the two groups. Therefore, the FACT-Cog total score for cognitive function from the above four subscale scores was significantly improved in the CACT+A group (Z = 3.026, P = .002), whereas this improvement was not found in the CACT group.

| Participation in everyday activities
As seen in Table 3, MOHOST subscale scores for process skills were significantly increased from pre to posttest in both groups (ps < .05), but there was not significant difference between the two groups. For the MOHOST subscale scores for pattern of occupation and environment factors, the change scores were significantly lower in the CACT +A group as compared to the CACT group, suggesting that the CACT +A treatment did not improve participation in everyday activities.

| Qualitative results
Five themes and their subthemes emerged from the data and generally reflected the focus of the questions (see Table 5). Themes were To analyze the qualitative data with the quantitative data, qualitative data were transformed into quantitative data, or quantized. 29 Code counting was the method used in quantizing the qualitative data (see Table 5). In response to question 1 as to changes in memory, con-   The computer-assisted cognitive exercises in our study did include those to improve working memory but all exercises were run for only up to 3 minutes per day. The primary author and colleague did not previously show improvement in working memory with CACT 12 and nor did Kesler and colleagues. 11 The authors have no plausible reason that the CACT significantly improved in digit span length but the CACT+A group did not.

| CONCLUSION
By using an embedded design in this mixed methods study, we were able to report results of quantitative outcomes measures and provide participants' perspectives on their performance in the study with the qualitative interview results. Notwithstanding study limitations, BCS who had the CACT+A program self-reported significantly improved perceived cognitive function and expressed experiencing greater ratios of improvements for memory and QOL. CACT+A is an auspicious intervention option for BCS who self-report cognitive issues. It is convenient to participate in at home and allows BCS to safely selfisolate if need be.