Breath-focused mindfulness and compassion training in parent-child dyads: a pilot intervention study

Background: Depression in children is a concerning societal issue, and can be associated with poor academic performance, school dropout and poor overall quality of life. Additionally, child depression is often associated with parallel stress and depression in parents.

Objective: This scenario highlights the urgent need for development and implementation of accessible and scalable solutions that may co-benefit child and parent mental well-being.

Methods: This pilot study introduced "Cooperative Compassion" (CoCo), a parent-child co-training digital application aimed at promoting mindfulness and compassion through brief, performance-adaptive sessions. A community sample of 24 parent-child dyads (children’s mean age: 9.5 years, 14 females/10 males, 11 Caucasian/5 Asian/7 mixed race/1 other race; parents’ mean age: 44.5 years, 20 females/4 males, 14 Caucasian/8 Asian/2 mixed race) of high average affluence socio-economic scores participated in the study. These parent-child dyads completed 30 sessions of CoCo training over three months with baseline and post-intervention assessments occurring within 2 weeks of training initiation/completion, respectively.

Results: The program was feasible, with 80% of families completing over 90% of sessions and providing positive feedback. Mental health assessments showed a non-significant effect in the expected direction in children’s depression scores (Cohen’s d=-0.19, 95% Confidence Interval (95% CI) [-8.89 to 1.74], P=.07) and significant reductions in parental stress (d=-0.41, 95% CI [-2.63 to -0.16], P=.02), anxiety (d=-0.47, 95%  CI [-2.67 to -0.20], P=.02), and depression (d=-0.50, 95% CI [-3.25 to -0.08], P=.03), with sustained benefits at the 3-month follow-up. Parental mindfulness improvements were correlated with stress reduction (rho=-0.45, P=.03). On an emotion bias task utilized as an objective assessment of cognition, children demonstrated improved processing speed post-intervention (d=0.54, 95% CI [0.011 to 0.088], P=.005), and a marginal improvement was also observed in parents (d= 0.19, 95% CI [-0.006  to 0.031], P=.05). Cortical source imaging of EEG recordings acquired simultaneous to an attention-to-breathing assessment showed significant reduction in task-related default mode network (DMN) activity (d=−0.53; 95% CI −0.0010 to −0.0003; P=.001).

Conclusions: Post CoCo intervention decrease in DMN activity on the attention-to-breath task in parent-child dyads may be indicative of cortical plasticity reflecting reduced mind-wandering and thereby, enhanced focus after training. The current promising behavioral and cognitive results suggest the need for a larger sample size and a randomized controlled study design. Overall, these findings highlight the potential for brief, digital mindfulness and compassion co-training to improve family mental health and well-being.

Methods
Participants. A total of 24 parent-child dyads participated in the current study. Dyads were recruited from local schools and university-affiliated pediatric clinics in the San Diego area through flyer advertisements and clinic referrals, respectively. Signed informed consent was obtained from the parents and signed assent was obtained from the children for study participation following the guidelines outlined in the Declaration of Helsinki. The study protocol was approved by the institutional review board of the University of California San Diego (protocol #180140). The study data were de-identified by taking out protected health information such as names, dates, and place of birth or anything else that is listed by Health Insurance Portability and Accountability Act (HIPPA) compliance. All the parent-child dyads were paid $250 for completing all assessments and digital training.

Also, all children and parents were right-handed and had normal or corrected-to-normal vision. Children were assessed on the Child Depression Index (CDI (Kovacs, 2012)) and study inclusion was based on CDI scores in the average and above-average range (T-score>40). Exclusion was based on any self-reported severe illness for parent or child that would not allow time for study participation. All parents were healthy and did not report any current diagnoses or medications. 

Sample Size and Power. This single-arm study was powered to detect medium effect size (Cohen d>0.5) comparing pre- versus post-intervention differences at β power of 0.8 and α level of 0.05 for each assessment measure. Effect sizes were calculated a priori using the G*Power software (Faul et al., 2009). 

Feasibility. We assessed intervention feasibility by monitoring the total number of assigned intervention sessions completed. Additionally, parents completed a feasibility survey at end of study that we have standardized in previous digital training studies (Mishra et al., 2016). The survey queried 16 questions about the training as elaborated in the Results (Table 2) and each question required a response on a seven-point Likert scale. The Cronbach α measure of reliability for the training survey was high (α = 0.92).

Assessments. Each parent-child dyad made two visits (baseline [pre-intervention] and three-months later [post-intervention]) to the Neural Engineering and Translational Labs (NEATLabs) and participated in behavioral and neurocognitive assessments. Behavioral assessments were also completed online at a 3-month follow-up after intervention completion.