Participants included K-12 (kindergarten through grade 12) educators in schools in the states of New York, New Jersey, Connecticut, and Pennsylvania, given that the impact of COVID-19 was similar among these regions (i.e., containment strategies, case counts). Subjects were recruited in three waves (from June 2020 to July 2020) through flyers distributed to area school administration (e.g., principals, assistant principals) and social media (i.e., Facebook) advertisements. Interested individuals with a physical health condition that may contraindicate vigorous exercise (e.g., history of heart disease, stroke, cardiac arrythmia, uncontrolled asthma, severe joint problems) were excluded from study participation. A computer-generated randomization sequence was used to assign participants to intervention (MAP Training) and waitlist control (No Training) groups using a ratio of 4:1 to obtain a sufficient sample size to test for treatment effects in the MAP Training group. One subject who expressed interest in participating in the study but was unable to attend the MAP Training sessions was thus assigned to the No Training group. The protocol was approved by the Rutgers IRB (Pro2020001365) and electronic informed consent was obtained for each subject prior to participation and reaffirmed at each assessment timepoint.

Mental and physical Training combines mental training with meditation and physical training with aerobic exercise (Shors et al., 2014; see Figure 1). This program was delivered online via the Zoom platform. First, participants were presented with a “brain bit,” which was a short piece of information about the brain to keep them engaged and motivated. Then, they watched and listened as the facilitator instructed them on how to set-up the meditation activities and then engaged in the activities along with the participants. The mental training component consisted of 20-min of a focused-attention (FA) meditation while sitting in silence, followed by 10-min of a walking meditation, again in silence. During the FA meditation, participants were instructed to breathe naturally while bringing their full attention to their breath. They were told to notice the short space between the out-breath and the in-breath (i.e., SIT; Figure 1). Participants were instructed to count the space between each breath, beginning with one and continuing until they lost count. Should their attention wander, they should acknowledge their thoughts without judgement, and return their attention to counting the space between each breath, beginning again with one. A timer was set to ring after 20 min, at which point participants were told to stretch out their legs before standing up. Once they felt ready, they were asked to stand up for the 10-min of walking meditation. During this part of the intervention, participants were instructed to clasp their hands loosely behind their backs and maintain their gaze at the floor ∼3 feet in front of them. They were to focus their attention on their feet while they walked a circular path at a very slow pace, noticing how the weight of the body changes with each step and meanwhile noticing how the bottom of the feet touch the floor (i.e., WALK; Figure 1). As during sitting meditation, participants were instructed to maintain attention on their feet as they walk until they lost concentration, at which time they were to recognize that they have lost their focus of attention and return it to the feet. Again, a timer was set to ring after 10 min. Next, the participants prepared themselves for the physical training component, which consisted of 30-min of moderate intensity aerobic exercise (i.e., SWEAT; Figure 1). Participants began the exercise component with a 5-min warm-up. Next, they were led through a choreographed aerobic exercise routine to popular music. Each session incorporated 9–10 tracks which were rotated in and out each week. The session concluded with a 5-min cool down. Each session was approximately 1 h.

Before training, participants were instructed to take their own heart rate by pressing their finger against the side of their neck and then asked in all sessions, to gauge their heart rate to ensure, to the extent possible, that exercise was performed at a moderate level of intensity. Approximately 20 min into the physical exercise component of each MAP Training session, subjects were directed to count their pulse over a 10 s period, then multiply the value by six. The aerobic range is generally defined as lying between 60 and 80% of participants’ maximum which is calculated by subtracting their age from 220. For most participants, their aerobic range was greater than 90 beats per minute but less than 140.

Analyses were conducted using SPSS Version 27 (IBM Corp, 2020). Between group (MAP Training, No Training) differences on baseline characteristics (i.e., sociodemographic variables [age, race, ethnicity, sex, educational attainment], health history, teaching experience, engagement in mindfulness, and exercise activities) were examined with independent samples t-tests and Pearson’s Chi-square tests for independence, where appropriate. An a priori power analysis indicated that a total sample size of 99 was necessary to detect large (f = 0.50) within-between interaction effects with 90% power (α = 0.05). As a result, the target sample size was 80 for the MAP Training group and 20 for No Training group (4:1 ratio).

A repeated-measures multivariate analyses of variance (MANOVA) tested between-group differences in primary (i.e., psychosocial) and secondary (i.e., cognitive, health) outcomes at pre- vs. post-intervention. Data were assessed for multivariate normality, homogeneity of covariance matrices, and multicollinearity. Significant univariate and multivariate interactions were followed with post hoc analyses. Group differences on outcome measures at pre-intervention and post-intervention timepoints were assessed with independent samples t-tests, whereas within-group changes were tested with paired samples t-tests. In both sets of analyses, a False Discovery Rate (FDR; Benjamini and Hochberg, 1995) correction (α = 0.05) was applied. The effects of MAP Training on primary and secondary outcomes were tested with repeated-measures MANOVAs. Significant univariate and multivariate interactions were followed up with independent and paired samples t-tests with an FDR correction applied. Data were further explored with pairwise comparisons, corrected for multiplicity.

Of the 104 teachers recruited, 71 completed the initial baseline assessment and were included in the data analyses. Of these participants, 58 were randomly assigned to the MAP Training group and the remaining participants to the No Training group. One participant was not randomly assigned because they could not attend the live sessions (Figure 2).

Groups did not differ significantly on characteristics prior to training [e.g., age, sex, race, years teaching, experience with mindfulness (including meditation), participation in exercise]. Rates of attrition did not differ by group [χ²(1, N = 104) = 0.16, p > 0.05]. Engagement was assessed after each live session by asking for self-reported engagement and maximal heart rate. On average, participants assigned to the MAP training group engaged in at least four out of the six live sessions and completed approximately half of the six recorded sessions. Based on this information, we considered participants who attended at least four out of six live MAP training sessions (n = 35) as treatment-adherent. These treatment-adherent participants and No Training groups did not differ significantly with respect to their characteristics before training (e.g., age, sex, race, years teaching, experience with mindfulness, participation in exercise, etc.; see Table 1). Treatment-adherent and non-adherent participants did not differ on most baseline characteristics, although a greater proportion of teachers in the treatment-adherent group reported no previous experience with mindfulness programs [χ²(2, N = 58) = 9.09, p = 0.01].

A one-way repeated measures MANOVA tested for significant differences between treatment adherent MAP training participants and those who received no training. The multivariate group × timepoint interaction was not significant (p = 0.06). The sample size in the No Training group (n = 11) was less than the number of dependent variables in the analysis (n = 12), and thus the analysis may have been underpowered. Nevertheless, a series of one-way ANOVA’s on these variables revealed a significant group x timepoint interaction on multiple variables, indicating that the MAP group demonstrated greater change than the No Training group on each of the following outcomes: overall mood and well-being [MAP Health Questionnaire; F_(1,43) = 5.40, p = 0.02, η² = 0.11], depressive symptoms [PHQ-9; F_(1,43) = 6.34, p = 0.02, η² = 0.13], anxiety symptoms [GAD-7; F_(1,43) = 13.15, p < 0.01, η² = 0.23], secondary traumatic stress related to profession [ProQoL-Secondary Traumatic Stress; F_(1,43) = 7.74, p = 0.01, η² = 0.15], and perceived stress [PSS-10; F_(1,43) = 7.97, p = 0.01, η² = 0.16]. A series of independent samples t-tests compared groups on psychosocial outcomes at each timepoint, with no significant differences at baseline (ps > 0.05). Post-intervention scores were significantly different between groups with respect to anxiety [GAD-7; t(46) = −2.40, p = 0.02, d = −0.78] and professional quality of life related to experiences of secondary traumatic stress [ProQoL-Secondary Traumatic Stress; t(46) = −2.31, p = 0.03, d = −0.75], with participants in the MAP Training group reporting lower levels of anxiety and less work-related stress than the No Training group, although between-group differences were not signficant after adjusting for multiple comparisons.

In the treatment-adherent MAP Training group, scores after the intervention significantly improved on measures of mood and well-being [MAP Health Questionnaire; t(34) = 3.82, p < 0.01, d = 0.65], depressive symptoms [PHQ-9; t(33) = 3.82, p < 0.01, d = 0.66], anxiety symptoms [GAD-7; t(34) = 3.36, p < 0.01, d = 0.57], rumination [RRS; t(34) = 2.91, p = 0.01, d = 0.49], brooding [RRS-Brooding; t(34) = 2.54, p = 0.016, d = 0.43], depressive ruminations [RRS-Depression; t(34) = 3.20, p < 0.01, d = 0.54], secondary traumatic stress related to profession [ProQoL-Secondary Traumatic Stress; t(34) = 3.11, p = 0.01, d = 0.53], perceived stress [PSS; t(33) = 2.19, p = 0.04, d = 0.38], and self-compassion [SCS-SF; t(34) = −4.87, p < 0.001, d = −0.82; Figures 3–5]. Participants in the No Training group reported an increase in anxiety symptoms [GAD-7; t(12) = −2.50, p = 0.03] from pre-intervention to post-intervention, although the change was not significant after adjusting for multiple comparisons.

Multivariate analyses of variance was applied to identify between-group differences in pre-intervention and post-intervention scores on cognitive measures, with no multivariate interaction (group × timepoint; p > 0.05). A series of independent samples t-test with FDR correction examined between group differences on cognitive measures at pre-intervention and post-intervention timepoints, consistent with the study’s a priori hypotheses. Pre-intervention scores on subjective estimates of cognitive flexibility [CFI; t(46) = −2.73, p = 0.01, d = −0.89], the ability to generate alternatives [CFI-Alternatives; t(46) = −2.69, p = 0.01, d = −0.87], and cognitive control [CFI-Control; t(46) = −2.30, p = 0.03, d = −0.75] were higher in the No Training group relative to the MAP Training group, but the difference was not significant after applying the FDR correction. Additionally, there were no significant differences between groups on these measures after the intervention (ps > 0.05). However, within the MAP group, post-intervention ratings on subjective measures of executive functioning [ADEXI; t(34) = 3.35, p = 0.01, d = 0.57], working memory [ADEXI-WM; t(34) = 4.37, p < 0.001, d = 0.74], cognitive flexibility [CFI; t(33) = −2.39, p = 0.02, d = −0.41] and cognitive control [CFI-Control; t(33) = −2.86, p = 0.01, d = −0.49] were significantly improved from baseline (Figure 6). There were no significant changes across time for teachers assigned to the No Training group.

In general, subjective assessments of physical health did not change as a result of the intervention, although participants in the MAP Training group reported significantly fewer sleep disturbances [sPSQI-Sleep Disturbances; t(34) = 3.36, p = 0.01, d = 0.57] at intervention end (Figure 7).

Ruminations tend to be negative, about the past, and infused with some degree of blame or regret. After training, the K-12 teachers reported fewer of these thoughts, including brooding and depressive subtypes. There were no changes in reflective rumination, which is about the past and tends to be less negative. There are numerous theories and still some controversy surrounding rumination subtypes, but in general, depressive rumination is most tightly linked to changes in mood while brooding rumination appears to predict later depression (Treynor et al., 2003; Burwell and Shirk, 2007). In this study, intervention participants also reported fewer symptoms of depression. Thus, the effects of MAP Training on more detrimental aspects of rumination (i.e., brooding and depressive rumination) may be in part responsible for alleviating some of the depression reported by the teachers during the pandemic. It is also conceivable that this type of intervention, when practiced routinely, may help prevent some increases in depression that can arise during stressful life events.

In addition to depression, rumination is linked to other aspects of mental health and wellness, including anxiety, trauma-related cognitions, and even how someone interprets changes in their body. In a recent factor analytic study, ruminations accounted for much of the variance (>96%) in mental health outcomes acquired through many of the same self-report measures used here (Millon and Shors, 2021). Of course, there are similarities amongst the questions in these surveys and these similarities may account in part for the relationships. But nonetheless, these analyses have led us to suggest that rumination may serve as a proxy for overall mental health. All that being said, we do not know the neural or psychological mechanisms through which rumination may affect other health outcomes. A recent meta-analysis did identify neural networks that were especially engaged in people who are inclined to ruminate whereas other networks, especially those in the temporal cortex were less engaged (Zhou et al., 2020). Perhaps these later networks are becoming more engaged because of training. In theory, this would be consistent with the increase in executive function and cognitive flexibility reported here by those who engaged in MAP Training. Functional imaging studies are underway to test this hypothesis.

Stress has wide-spread effects on mental and physical health, many of which arise not only from exposure to the stressful life event itself, but also from one’s subjective appraisal of the experience (Lazarus and Folkman, 1984). Our results indicate that MAP Training lessened teachers’ perceived stress. This finding is especially meaningful given that teaching is recognized as a high-stress profession (Shernoff et al., 2011; Richards, 2012; Gallup, 2014; McCarthy, 2019), with job-stress increasing for many teachers after the onset of the COVID-19 pandemic. In addition to the direct experience of stress, teachers, like other helping professions, are susceptible to secondary traumatic stress (Hydon et al., 2015; Molnar et al., 2017). Secondary traumatic stress differs from vicarious trauma in that it involves emotional and behavioral reactions to secondary exposure to trauma as opposed to changes in cognitive schemas and beliefs (Jenkins and Baird, 2002). In this study, MAP Training led to a significant reduction in secondary traumatic stress. Therefore, MAP Training may lessen the risk and thus help prevent secondary traumatic stress and associated distress among teachers who may have been exposed to student trauma during the COVID-19 pandemic (i.e., Collin-Vézina et al., 2020; Kovler et al., 2021).

Despite reported improvements in mood, anxiety, and indicators of stress, we did not observe changes in teachers’ ratings of distress tolerance in this study. Interestingly, intolerance of uncertainty, a subcomponent of distress tolerance, has been identified as a unique predictor of future perceived stress (Bardeen et al., 2016). Thus, distress tolerance may have been somewhat impervious to the effects of the intervention given the general state of uncertainty and anxiety during the initial months of the COVID-19 pandemic (Twenge and Joiner, 2020). It is also important to note that distress tolerance is commonly conceptualized as a relatively stable, trait-like marker of psychopathology symptoms (Leyro et al., 2010). Moreover, some data suggest that standalone mindfulness interventions may lead to changes in behavioral distress (e.g., task persistence) but not perceived distress tolerance (e.g., Carpenter et al., 2019). Taken together, these data suggest that MAP Training may be effective in improving one’s current experience of distress as opposed to their perceived ability to relate to symptoms (i.e., distress tolerance).

Virtual delivery of MAP Training during the pandemic also had a positive impact on teachers’ subjective ratings of cognitive functioning. Treatment adherent participants reported significant improvements in working memory and executive functioning of medium effect size after 6 weeks of MAP Training. Additionally, participants who adhered to the program reported significant improvements in cognitive flexibility and control following training, whereas no such change was observed among waitlist control participants. The impact of MAP Training on self-reported cognition in this study is consistent with previous research on the cognitive effects of mindfulness meditation interventions (Chiesa et al., 2011) as well as considerable data demonstrating a positive link between exercise and brain function (i.e., Ludyga et al., 2020). And as noted, MAP Training has been associated with an increase in amplitude of the early components of the evoked response during the Flanker task, which engages neural processes related to executive function and cognitive control (Alderman et al., 2016).

Aerobic exercise alone has numerous effects on brain function, including increases in vascular growth, neurogenesis in the hippocampus, as well as the release of growth factors such as BDNF (i.e., van Praag et al., 1999; Piepmeier and Etnier, 2015; Phillips, 2017; Vivar and van Praag, 2017; Liu and Nusslock, 2018). Meditation has also been linked with increases in hippocampal volume among those who practice meditation regularly (Luders et al., 2013) and reductions in hippocampal atrophy in those with mild cognitive impairment (Wells et al., 2013), possibly through effects of synaptogenesis, angiogenesis and neurogenesis. In fact, MAP Training was developed for humans based on preclinical studies suggesting that mental training with effortful learning procedures increases neurogenesis in the adult hippocampus (Gould et al., 1999; Leuner et al., 2004; Curlik and Shors, 2011; Shors, 2021). Others have reported that a combination of spontaneous learning during environmental enrichment when preceded by aerobic exercise is especially neurogenic (Fabel et al., 2009). Therefore, the combination of mental (i.e., FA meditation) and physical (i.e., aerobic exercise) training coupled with enhanced mood may produce cognitive change in humans through mechanisms of hippocampal plasticity.

The combination of mental and physical training had a positive impact on teachers’ physical wellbeing, specifically on the quality of sleep. Participants reported fewer sleep disturbances including fewer nighttime awakenings, breathing difficulties and nightmares relative to their reports before training. There were no such changes in the waitlist control condition. Sleep quality is an important predictor of subjective well-being (Reid et al., 2006; Peach et al., 2016; Wickham et al., 2020) and demonstrates a bidirectional relationship with depression and anxiety symptoms (Alvaro et al., 2013). Given the increased prevalence of poor sleep quality during the COVID-19 pandemic (i.e., Gupta et al., 2020; Pinto et al., 2020; Hyun et al., 2021), interventions that improve sleep, even indirectly, are especially needed.

MAP Training does not necessarily impact mental health through direct changes in cardiovascular activity. For example, we observed no change in heart-rate variability and related measures of sympathetic nervous system activity after women with HIV completed 6 weeks of training, despite robust decreases in ruminative and trauma-related thoughts (Millon et al., 2022). However, training did increase the volume of oxygen consumed in women with physical complaints such as addiction and malnutrition (Shors et al., 2014). In contrast, the current participants were relatively young high-functioning adults. It is likely that individual differences in physical health prior to training are important (i.e., the potential range for change) as well as the length of the intervention, which is relatively short per session (1 h) and over its course (6 weeks).

There are several limitations to this research. First, our results were limited by a less than desirable sample size and suboptimal adherence, with most participants completing about four out of six of the live sessions, and half of the weekly recorded sessions. High rates of attrition are commonly observed in studies of virtual and web-based interventions (Eysenbach, 2005; Melville et al., 2010). In this study, dropout rates may have been exacerbated by an overall increase in anxiety, depression, and stress in the general population during the COVID-19 pandemic (e.g., Xiong et al., 2020), and especially among teachers (e.g., Aperribai et al., 2020; Anderson et al., 2021; Ozamiz-Etxebarria et al., 2021). Indeed, data from reviews indicate that the COVID-19 pandemic has resulted in substantial declines in participant enrollment in clinical trials and research studies (Sathian et al., 2020). Additionally, we began recruitment during the beginning of summer when most teachers were out or soon to be out of the classroom. And thus, burnout and workplace fatigue may have lessened their willingness to participate (Pressley, 2021). Because the conditions of the pandemic were constantly changing, we could not continue to recruit once they had returned in the fall to the classroom. Yet, despite these restrictions, we observed significant and positive effects of MAP Training on subjective estimates of mental health and well-being.

The MAP Training intervention was delivered virtually to accommodate stay-at-home orders and remote working conditions during the COVID-19 pandemic. Importantly, the teachers reported a high level of engagement and their average heart rate during the physical training was 130 beats per minute, suggesting that they were exercising at an intensity consistent with aerobic exercise. Nevertheless, there were certain limitations to the virtual format. For example, the participants’ cameras were disabled during the live sessions to protect their privacy. (And several participants stated beforehand that they did not wish to turn on the camera). Therefore, it is not known whether participants were fully adhering to the MAP Training intervention. Additionally, variability in internet and technology literacy and quality may have interfered with their engagement. These potential barriers have been documented in prior studies of virtual interventions (e.g., Borghouts et al., 2021) and have yet to be adequately addressed.

Finally, we recruited teachers within the local tri-state area because of similarities in the impact of COVID-19 in this region. At the time of the study, COVID-19 cases were among the highest in the country and residents were facing statewide travel restrictions, mask mandates, limits on social gatherings, and supply shortages. Thus, we cannot be certain that our findings would generalize to other populations of teachers within or outside of the United States.

Overall, findings from this pilot study suggest that 6 weeks of virtual MAP Training can lead to positive changes in select measures of mental health, especially those related to mood, negative thinking, and overall well-being. Even prior to the COVID-19 pandemic teachers were affected by high levels of occupational stress (Gallup, 2014), which has been associated with poor mental health outcomes, including depression and anxiety (Besse et al., 2015; Jones-Rincon and Howard, 2019). Moreover, workplace burnout has been closely linked with depression (Ahola et al., 2014), with some studies suggesting significant overlap between the two constructs (Schonfeld and Bianchi, 2016). Therefore, interventions that target psychological distress, such as MAP Training, may help to reduce burnout and subsequent rates of attrition among teachers, even under normal working conditions. The mental and physical health benefits of meditation and aerobic exercise have been demonstrated independently in a variety of populations (e.g., Warburton, 2006; Gillison et al., 2009; Chiesa et al., 2011; Keng et al., 2011; Kvam et al., 2016; Stubbs et al., 2017; Creswell et al., 2019), including teachers (e.g., Crain et al., 2017; Klingbeil and Renshaw, 2018; Zarate et al., 2019; Abós et al., 2021). The present results highlight the potential benefit of combining these activities together in one intervention to alleviate stress and promote well-being in teachers. They further suggest that the virtual delivery of interventions such as MAP Training are effective in improving mental health and mitigating the impact of stressful life events, such as occurred during the coronavirus pandemic.