Exercise Promotes Neuroplasticity in Both Healthy and Depressed Brains: An fMRI Pilot Study
Memory impairments are a frequently reported cognitive symptom in people suffering from major depressive disorder (MDD) and often persist despite antidepressant therapy. Neuroimaging studies have identified abnormal hippocampal activity during memory processes in MDD. Exercise as an ad-on treatment for MDD is a promising therapeutic strategy shown to improve mood, cognitive function, and neural structure and function. To advance our understanding of how exercise impacts neural function in MDD, we must also understand how exercise impacts healthy individuals without MDD. This pilot study used a subsequent memory paradigm to investigate the effects of an eight-week exercise intervention on hippocampal function in low-active healthy (n = 8) and low-active MDD (n = 8) individuals. Results showed a marked improvement in depression scores for the MDD group (p < 0.0001) and no change in memory performance for either group (p > 0.05). Functional imaging results showed a marginally significant decrease in hippocampal activity in both groups following the exercise intervention. Our whole brain analysis collapsed across groups revealed a similar deactivation pattern across several memory-associated regions. These results suggest that exercise may enhance neural efficiency in low-fit individuals while still resulting in a substantially greater mood effect for those suffering from MDD. This trial is registered with clinical trials.gov NCT03191994.
Memory impairment is the most frequently reported cognitive symptom in people suffering from major depressive disorder (MDD) and often persists as a residual symptom following antidepressant therapy [1, 2]. Although the neural underpinnings of impaired memory in MDD remain unclear, research suggests that the hippocampus, which plays a critical role in the formation of new memories, also plays a role in the pathogenesis of MDD. To date, cognitive literature has presented mixed findings in terms of the type, severity, and specificity of memory deficits in people with MDD, although the plurality of data has suggested an impairment in episodic (autobiographical) memory with a sparing of semantic (general knowledge) memory and short-term memory [3–8]. The hippocampus has been shown to play an essential role in the encoding of episodic memories [5, 9–13], and pathologies associated with this neural structure may underlie the episodic memory impairments observed in MDD populations. The relationship between MDD, memory impairments, and hippocampal structure and function is based on converging lines of research from animal studies, neuroimaging, neuropsychology, and postmortem investigations which have all shown hippocampal abnormalities at the structural, functional, and cellular level. Structural brain imaging studies have shown robust hippocampal volume reductions particularly in persistent and early onset MDD [14–16]. Functional neuroimaging studies have found that both the memory encoding and retrieval processes within the hippocampus are to be impaired in MDD [4, 6, 17–19]. Neuropathological evidence from animal models of depression and postmortem studies in depressed humans have revealed cellular abnormalities in the hippocampus such as dendritic atrophy and reduced neuron and glial densities [20–24].
Despite many efforts to develop effective antidepressant therapies, MDD remains a severely undertreated disorder in the primary care setting leaving more than half of individuals plagued with symptoms [25–28]. Exercise as an add-on to conventional antidepressant therapies is a promising treatment strategy for MDD. It is well established that exercise is efficacious in treating mild to moderate depression with response rates comparable to mainstream therapies such as antidepressant medication and cognitive behavioral therapy [29–34]. However, there is a lack of understanding of the neurobiological mechanisms that underlie or mediate the antidepressant effects of exercise. It is well established that exercise facilitates neuroplasticity [35–37]. To date, much of our understanding of how exercise facilitates neural and cognitive plasticity has come from the extensive animal literature. For instance, rodent studies have shown that exercise increases new cells in the dentate gyrus of the hippocampus, and this is associated with improved learning and spatial memory [36, 38–40]. Further evidence has shown that exercise increases synaptogenesis  and angiogenesis  and improves dendritic morphology in the hippocampus [43, 44]. However, the effects of exercise on brain structure and function in humans have been more equivocal. In elderly populations, aerobics exercise training has been shown to improve spatial memory , executive function [46, 47], and short-term memory ; however, others observed no benefits [49–52]. Neuroimaging studies have found that aerobics exercise reverses age-associated brain volume loss in the prefrontal and temporal cortices [53, 54], improves functional connectivity in the default mode and frontal executive networks , and increases hippocampal volume in schizophrenics . To date, the literature examining the effects of chronic exercise on neural function in both healthy and MDD populations remains scant. To capitalize on the full treatment potential of exercise in MDD populations, we must also understand the relationship between cardiovascular fitness, neural function, and cognitive performance in healthy individuals in order to identify neural mechanisms specific to MDD.
To our knowledge, this is the first study using a subsequent memory paradigm to determine the effects of an eight-week exercise prescription on the functional integrity of the hippocampus in low-active patients with MDD and low-active healthy individuals. The aim of this pilot study was twofold: (1) using fMRI to examine changes in hippocampal function following an exercise intervention, and (2) to conduct an exploratory whole brain analysis to determine how exercise affects overall brain activity.
2. Materials and Methods
Eight patients (mean age = 37.25, SD = 8.00; 7 females) with comorbid MDD and anxiety were recruited from an outpatient Mental Health Day Treatment Program at a local hospital in Oshawa, Ontario Canada. Eight healthy participants (mean age = 20.63, SD = 1.19; 4 females) with no history of mental health illness or neurological disease were recruited from a local university in Oshawa, Ontario Canada. Depressed patients had a diagnosis of MDD according to an unstructured clinical interview by hospital psychiatrists based on Diagnostic and Statistical Manual of Mental Disorders—Fourth Edition (DSM-IV-TR)  criteria, no coexisting DSM-V Axis I disorders apart from anxiety, and a score ≥20 on the Beck Depression Inventory—Second Edition (BDI-II) , and their pharmacological treatment was stabilized a minimum of six weeks prior to study enrolment. In order to be considered eligible for the study, participants needed to indicate that they exercised less than 20 minutes, three times per week. Both groups were also safety screened for MRI and screened with the Physical Activity Readiness Questionnaire (PAR-Q) to ensure they had no medical contraindications to exercise. All participants provided written consent.
2.2. Psychometric Evaluation
Participants completed the Montreal Cognitive Assessment (MoCA) which is a brief neurocognitive tool with high sensitivity for screening patients with mild cognitive impairment. This cognitive assessment was performed to identify participants who may have difficulty performing the associative memory task. The internal consistency of the MoCA is good, with a coefficient alpha of 0.83 . Depression was measured using the Beck Depression Inventory (BDI-II)  which is one of the most widely used self-reported instrument capable of measuring depression severity ranging from not depressed to severely depressed. The BDI–II demonstrated excellent internal consistency, with a coefficient alpha of 0.91 . All measures were performed before and after the eight-week exercise intervention.
2.3. Fitness Assessment
Cardiorespiratory fitness was measured before and after the exercise intervention using the YMCA cycle ergometer protocol recommended by the American College of Sports Medicine [60–62]. The YMCA cycle ergometer protocol is an indirect submaximal exercise test used to estimate maximal oxygen consumption (VO2max) from heart rate (HR) measurements. The protocol consists of two or more consecutive 3-minute stages at a given workload. The objective was to elevate the participant's HR to a target zone to approximately 85% of the age-predicted maximum HR for two consecutive stages. The initial workload consisted of a 25-watt workload at a cadence of 50 revolutions per minute. HR was measured and recorded using the radial pulse method during the final 15 seconds of each minute, which determined the workload of subsequent stages indicated by the YMCA protocol. Once a steady state HR (two successive measures that differ from <5 bpm) was within 10 bpm of the 85% age-predicted maximum HR, the test was complete. VO2max was estimated using an equation that includes workload, body mass, and derived constants.
2.4. Exercise Intervention
Participants performed an individualized eight-week exercise program consisting of three weekly sessions (described below). Exercise sessions were performed alone, on nonconsecutive days, and each session was supervised by a qualified exercise professional to increase compliance . Attendance was recorded and all participants completed >80% of the exercise sessions.
The exercise prescription was based on international recommendation to obtain at least 150 minutes per week of moderate to vigorous intensity aerobics exercise and to perform strengthening activities twice per week, for developing and maintaining cardiorespiratory, musculoskeletal, and neuromotor fitness in healthy adults [63, 64]. This minimum-effective dose of exercise was prescribed to encourage better compliance since people with depression, and are low active, tend to be less motivated . Research has also shown that combining aerobics with strength training improves depression and cognitive function such as attention, processing speed, executive function, and memory performance more than aerobics exercise alone [51, 66].
2.4.1. Resistance Sessions
Resistance sessions were completed twice per week and incorporated a whole-body exercise prescription using larger muscle groups. Each session included eight resistance exercises using both resistance training machines and free weights. Initial workloads were approximately 95% of the 10 repetition maximum to ensure proper form. Exercises were performed in two or three supersets (one set of each exercise with no rest between sets) with an 8–12 repetition range in order to decrease rest times and to maintain target HR. Workload was increased approximately 5% once participants were able to complete three sets of 12 repetitions with proper form. Specific exercises were changed every four weeks; however, they targeted the same muscle groups. Resistance exercises included variations of the chest press, pull downs, triceps extension, biceps curl, shoulder press, leg press, leg extensions, leg curls, squats, split squats, calf raises, and abdominal exercises. During each session, HR was monitored to ensure that the participant maintained a HR of between 60–80% of their age-predicted maximum HR. Each session began with a 5-minute aerobics warmup and ended with 15 minutes of aerobics activity that was performed on either the treadmill, stationary bike, or elliptical trainer.
2.4.2. Aerobics Session
Participants completed an aerobics-only session once per week. They were given the choice to perform their aerobics activity on either the treadmill, stationary bike, or elliptical trainer. The aerobics session workloads were determined by HR response and increased by five-minute increments over the eight weeks reaching a maximum of 60 minutes per session. HR was monitored throughout the session to ensure the participant was in the target HR range.
2.5. Statistical Analysis
Statistical analyses were performed using GraphPad Prism software, version 6.0 data. Data are presented as mean (standard deviation (SD)). P values less than 0.05 were considered significant. Differences in baseline variables between groups were tested using a two-tailed Student's t-test and chi-square test for gender distribution. Within group differences for pre-post-BDI, MoCA, BMI, and VO2max were tested using a paired t-test. A two-way repeated measures analysis of variance (ANOVA) was used to determine any group × time of interactions and to compare the changes between the two treatment groups. Cohen's d was used to represent the effect size within each group. For between group effect sizes, we used dppc2  which uses the difference between Hedge's g of two different treatment groups in pre-post research designs.
2.6. Associative Memory Task
To evaluate the encoding and retrieval processes of memory, fMRI studies frequently use a recognition memory paradigm that consists of an “encoding” and “recall” phase [5, 68]. Associative memory refers to memory for the relationships between memoranda rather than memory for objects themselves [69–71]. The role of the hippocampus in memory formation has also been specifically linked to associative memory . A specific version of an associative paradigm  using face-name pairing known to reliably activate the hippocampus during the successful encoding event and sensitive enough to detect hippocampal dysregulation in a MDD sample  was used to investigate activation patterns of the hippocampus during the encoding process inside the MRI scanner (see Figure 1). During the encoding phase inside the MRI, participants were presented with face-name pairs and used a response box provided to indicate if the name suited the face. The retrieval phase was performed after the MRI scan. Participants were presented with a face and two names and instructed to indicate which name was paired with that face during the encoding phase. Participants also rated the confidence of their responses. Trials during the encoding phase were then reclassified based on the responses during the retrieval phase into correct (the participant selected the correct name for the face and indicated a high confidence in their response, suggesting that the face-name association was successfully encoded), guesses (a correct selection with low confidence), or incorrect (the wrong name was selected).
2.7. fMRI Scanning Parameters
Participants were scanned on a 3-Tesla Tim trio MRI scanner equipped with a 32-channel phased array head coil. E-prime software version 2.0 (Psychology Software Tools) was used to present stimuli on a rear-projection system (Avotec, Inc., Stuart, FL) in two separate nine-minute functional runs. To obtain optimal hippocampal resolution, all scans were acquired in the oblique coronal plane perpendicular to the long axis of the hippocampus to maximize the anatomic delineation. A total of 416 functional scans were acquired with a T2∗-weighted gradient echo planar imaging sequence (TR =2500 ms, TE = 27 ms, FOV = 192 mm, 3 mm × 3 mm × 3 mm, and flip angle = 70°; in-plane resolution = 3 mm × 3 mm; and 50 slices with 3.5 mm slice thickness). The first 4 volumes of each run were discarded to allow for T1 equilibration. The anatomical scan lasted six minutes and was acquired with a T1 MPRAGE imaging sequence (TR = 2000 ms, TE = 2.63 ms, FOV 256 mm, 1 mm × 1 mm × 1 mm voxels, and flip angle = 9°).
2.8. fMRI Analysis
Image preprocessing was performed using SPM12 methods (Statistical Parametric Mapping, Wellcome Department of Cognitive Neurology, London, UK: http://www.fil.ion.ucl.ac.uk/spm) within MatLab 8.3 (The MathWorks Inc., MA). Individual functional images were slice time corrected and realigned to the first image in the series to correct for motion. The EPI images were coregistered to the T1, and segmentation was applied to the T1 anatomical images to extract grey matter, white matter, and CSF masks and calculate a deformation field to transform the data into MNI space. All EPI images were then spatially normalized to the ICBM template using the deformation field, resampled to 3 × 3 × 3, and smoothed using a 6 mm full-width-half maximum isotropic Gaussian filter. General linear model (GLM) was performed at the single-subject level and statistical contrasts were created modeling the hemodynamic response function (HRF) of remembered items with high confidence (correct), remembered items with low confidence (lucky guess), and incorrect trials (incorrect). Six head motion parameters (three rigid body translations and three rotations) were included in the model to reduce the potential effects of motion. Second-level random effects analysis was performed using the contrast of t-test of correct > incorrect. Correct retrieval requires well encoding of the items. As such, this contrast which differentiates between poorly and well-encoded trials, known as the subsequent memory effects, is a powerful tool for examining successful memory encoding in the brain [73, 74]. As our primary hypothesis was related to activity in the hippocampus, a hippocampal mask was defined using the automated anatomical labeling (AAL) atlas. Significant clusters from an independent sample t-test within the hippocampus (p < 0.01 uncorrected, 10 voxels, for this a priori ROI-defining analysis only) for correct > incorrect at baseline were used as an ROI to extract contrast beta values for correct > incorrect in pre- and postscans for each participant. The average beta values from each ROI were imported into SPSS version 20, and a 2 × 2 repeated measures ANOVA (group × time) was run. We then conducted an exploratory whole brain analysis to determine if other brain regions showed task-related effects. For our whole brain analysis, significant clusters were defined as 20 contiguous voxels (180 mm3) with p < 0.005 uncorrected. A 2 × 2 repeated measures ANOVA (group × time) was run on β values for the correct > incorrect contrast in order to identify regions in which there was a difference in pre-post changes across groups. Additionally, an exploratory analysis was run examining a group × time interaction across the whole brain.
3.1. Baseline Characteristics
At baseline, there were no significant differences between groups for BMI, VO2max, and MoCA scores (all p > 0.05). BDI scores for the MDD group scores indicated severe depression while the healthy group BDI scores indicated no depression (p < 0.0001). The depressed group was also older than the healthy group (p < 0.0001). See Table 1.
3.2. Psychometric, Memory, and Fitness Results
A 2 × 2 repeated measures ANOVA revealed a group × time interaction for BDI scores (f (1,15) = 30.42, p < 0.0001) indicating that the MDD group had a greater decrease in depression scores pre-post. There were no significant changes in BMI, MoCA scores, or performance on the associative memory task (p > 0.05) for either group pre-post. Although baseline memory scores between groups were not significantly different (p = 0.477), our results showed that the MDD group performed more poorly on the associative memory task compared to the healthy group (71.48% versus 75.32%) indicating likely memory impairments in the MDD group. One MDD (n = 1) participant discontinued baseline VO2max testing due to exhaustion and was excluded from VO2max analysis. Baseline VO2max scores revealed that one MDD participant (n = 1) was in the good health benefit rating zone, and the remaining participants (n = 14) were in the poor health benefit rating zone based on the Canadian Society for Exercise Physiology guidelines . The healthy group showed a 47% increase in VO2max that was significant (p = 0.014) while the MDD group showed a marginally significant increase of 31% (p = 0.073). There was no significant difference in VO2max between groups (p = 0.661). These improvements in VO2max suggest that the exercise intervention was successful at improving cardiorespiratory fitness (see Table 2).
3.3. fMRI Results
Using a main effect contrast of correct > incorrect, collapsing across groups, at baseline, we identified active voxels in the hippocampus and created ROIs for the right and left hippocampus (see Figure 2(a)). A repeated measures ANOVA examining group × time using β values for the correct > incorrect contrasts for pre and post revealed a marginal main effect of time (f (1,15) = 3.3, p = 0.09), no main effect of the group (f (1,15) = 0.005, p = 0.957) or group × time interaction (f (1,15) = 0.165, p = 0.69). This marginal main effect of time is driven by a decrease in the correct > incorrect contrast in the hippocampus indicating that there was a decrease in hippocampal activity during successful encoding in both groups following the exercise intervention (see Figure 2(b)).
The exploratory whole brain analysis did not reveal any clusters in which there were group differences in the pre-post changes following exercise or a main effect of group differences. Given the lack of interaction of the main effects of the group or group × time interaction, a post hoc whole brain analysis was run in which both groups were collapsed. Given the relatively small sample size, this analysis maximizes the power to detect changes in brain activity following the exercise intervention that are common to both healthy and MDD, using the more powerful paired sample t-test. Changes in neural activity in the correct > incorrect contrast were compared from the pretreatment to posttreatment MRI scans. Decreases in activity following exercise were noted in several regions (see Figure 3 and Table 3). Regions included a larger cluster in the left posterior insula and smaller clusters in the medial superior frontal/mid cingulate and postcentral superior parietal gyrus.
In order to examine regions in which changes in neural activity were related to changes in BDI score, an additional contrast was run, regressing change in BDI against the paired t-test described above. Again, in order to maximize power, the healthy and MDD cases were both included in this analysis. The justifications for including both groups are as follows: firstly, although the healthy group was not clinically depressed, there was some pre-post reduction in BDI scores for the healthy group, and secondly, since cases with depression tended to have a larger decrease in BDI, this analysis may be more sensitive to group × time effects, reflected as BDI changes, while also better reflecting areas in which the pre-post differences were behaviorally meaningful. The regression against BDI for pre-post changes found a negative relationship between changes in depression scores and activation in the right occipital, left occipital/fusiform, and left precentral gyrus (see Figure 4 and Table 4).
This small fMRI pilot study used a subsequent memory paradigm to investigate the effects of an eight-week structured, supervised exercise intervention on hippocampal function and overall brain activity in low-active patients with MDD and low-active healthy individuals. The current study yielded two main findings. First, our ROI analysis of the hippocampus showed a marginal decrease in activation for both groups pre-post exercise. Although this decrease in hippocampal activation was only marginally significant, a deactivation pattern was present in both groups and was consistent across other memory-related brain regions noted in the whole brain analysis. These data provide the first evidence that improved cardiovascular fitness, following eight weeks of the minimum recommended dose of exercise, affects neural function alike in healthy and MDD brains. The overall deactivation pattern that we observed in the hippocampus and several other brain regions despite similar memory performance pre-post suggests increased cortical inhibition that attenuated neural activity in a subset of brain regions known to inhibit memory encoding and/or an increase in neural network efficiency during the memory encoding process. Second, our study showed that exercise had a robust antidepressant effect on the MDD group who went from the severe to mild depression range, providing additional support to the growing body of literature that exercise is an effective adjunctive therapy for MDD .
A common theme in the neurocognitive literature is that brain activity for remembered items is greater than brain activity for forgotten items, as this suggests successful memory encoding [17, 76–79]. However, neuroimaging studies employing a subsequent memory design have identified a negative relationship between remembered items and neural activity in brain regions such as the insula and the supramarginal gyrus, and hyperactivity in these regions may be detrimental to new memory formation [76, 80–82]. A candidate mechanism for the decrease in neural activity that we observed following the exercise intervention may be a modulation in the main inhibitory neurotransmitter γ-aminobutyric acid (GABA). Cortical inhibition, mediated by GABA via cortical interneurons, is an essential mechanism that eliminates task-irrelevant distractors that increase neural noise, which negatively affects attention for task demands. Inhibitory pathways consisting of GABAergic projections between the thalamus and cortex provide a mechanism that may eliminate task-irrelevant distractors by suppressing irrelevant sensory inputs early in sensory processing [83, 84]. A plethora of evidence has identified GABA deficits in MDD, and it has been postulated that GABAergic dysregulation may play a significant role in the pathogenesis of the disorder [85–88]. For example, neuroimaging studies have identified GABA deficits in the dorsolateral prefrontal and occipital cortex in depressed individuals [89–91]. Histopathological studies of postmortem tissue from MDD brains have revealed a reduction in both the density and size of GABAergic neurons in the prefrontal and occipital cortex that conceivably underlie the low levels of GABA seen in neuroimaging studies [92, 93]. Research has shown that exercise may facilitate cortical inhibition by regulating the interplay between glutamatergic excitatory neurons and GABAergic inhibitory interneurons. In mice, running engaged inhibitory mechanisms in the hippocampus through an increased expression of vesicular GABA transporter and extracellular GABA release that was also associated with improved anxiety regulation . In humans with early Parkinson's disease, a neurophysiological study used transcranial magnetic stimulation (TMS) to examine cortical inhibition of the primary motor cortex (M1) following an eight-week, high-intensity aerobics exercise intervention. In addition to improving clinical symptoms, the exercise intervention normalized corticomotor excitability through an increase in GABA-mediated cortical inhibition . Nonetheless, literature supporting the role of exercise in normalizing cortical inhibition via the GABAergic system remains sparse.
Our observed decrease in brain activity during successful memory encoding pre-post the exercise intervention also suggests lowered demands on neural networks and increased neural processing efficiency. Our results provide additional support to a recent body of literature, which postulates that exercise increases neural efficiency. In children, an eight-month aerobics exercise program was associated with decreased activity in several brain regions during an antisaccade task alongside improvements in performance . In elderly adults, a 12-week aerobics exercise program was associated with decreased prefrontal activation despite improvements in visual short-term memory . A similar study conducted in elderly adults with mild cognitive impairment found that 12 weeks of aerobics exercise decreased brain activity in 11 brain regions during memory retrieval despite improvements in memory performance . In adolescence, high-fit individuals showed a pattern of decreased activation in the hippocampus and right superior frontal gyrus combined with a deactivation in the default mode network (DMN) during the encoding of subsequently remembered items, that was absent in low-fit individuals . To determine if aerobics exercise influences learning and memory-associated neural circuitry, a group of researchers examined the brain activity in high-fit and low-fit adolescents during an SME paradigm. Despite comparable memory performance between the two groups, there were notable differences in memory-related and default mode (DMN) brain regions during encoding of successfully remembered word pairs versus forgotten word pairs. Results showed that high-fit individuals displayed a robust deactivation pattern in the DMN areas, such as the ventral medial prefrontal cortex and posterior cingulate cortex, which was absent in the low-fit group. The low-fit group also showed a greater bilateral hippocampal and right superior frontal gyrus activation during encoding of later remembered versus forgotten word pairs. Our results taken together with previous research suggest that improvements in aerobics fitness from the exercise intervention can promote neural processing efficiency during memory encoding processes.
Finally, the neurocognitive benefits associated with exercise may be attributed to increases in cerebral blood flow and neural growth factors, particularly brain-derived neurotrophic factor (BDNF), a key mediator of neuroplasticity in the brain [36, 37]. BDNF, a member of the neurotrophin family, upregulates neurogenesis, promotes neural survival, improves neural structure, and increases synaptic efficacy [100–103]. BDNF also modulates the formation and plasticity of GABAergic synapses and promotes maturation of GABAergic inhibitory networks [104–106]. Reduced BDNF levels are a consistent finding in animal models of depression , and administration of exogenous BDNF into the hippocampus is able to produce antidepressant behavioral responses comparable to antidepressant medications . Exercise is known to elevate BDNF production in the hippocampus [35, 109] and has been postulated as a leading candidate mechanism underlying the antidepressant effects of exercise [110, 111].
This pilot study has some limitations. First, the sample size used is rather small and therefore statistically underpowered. We did not age match our groups, which resulted in the MDD group being significantly older than the healthy group. Intrinsically, we wanted to compare MDD brains to young healthy brains with no history of mental health illness or other confounding comorbidities that increase with age and determine if exercise affects neural function in healthy populations who are low active. Also, we did not measure sedentary behavior time which has been shown to have deleterious health consequences independent of daily physical activity levels ; as a result, future work must consider sedentary behavior independent of physical activity levels and cardiorespiratory fitness. Another limitation is that our MDD samples were all medicated which may have also affected results. However, this is the typical patient seen in clinical practice, and in any real-world clinical intervention using exercise, the participants would likewise be similarly medicated. Next, even though our analysis collapsing across groups mediates some of the power issues for a transdiagnostic analysis of the effects of exercise on neural activity, the analysis was still underpowered to detect group × time interaction effects.
While we did not closely replicate the results of Fairhall et al., it should be noted that they used a contrast comparing correct trials to fixation, while we made use of a more standard subsequent memory contrast (correct > incorrect). Nevertheless, we did observe a decrease in the correct > incorrect contrast in the hippocampus indicating that there was a decrease in hippocampal activity during successful encoding in both groups following the exercise intervention. We did not however observe any group effects. Given the small sample size, we were likely underpowered to detect any subtle group effects, though it remains possible that the decreases in activity observed following the intervention are common across low-active individuals regardless of diagnostic status. In fact, it should also be noted there was even a small decrease in BDI scores amongst the healthy group. The physiological effects of exercise on the brain may be common amongst both healthy and MDD while still resulting in a substantially greater mood effect for those suffering from depression.
6. Future Research
This small pilot study demonstrates that eight weeks of the minimum recommended dose of exercise improved cardiorespiratory and significantly reduced depression severity in the MDD group. Importantly, we were able to demonstrate that combining the minimum recommended dose of exercise with conventional treatments was effective in treating the typical patient seen in the primary care setting who continues to experience severe depressive symptoms despite being treated with antidepressant medication. On the other hand, prescribing exercise to MDD patients presents many challenges to the practitioner since many patients lack motivation to initiate and maintain an exercise routine. Introducing patients to the minimum recommended dose of exercise as an add-on therapy may offer a practical approach for practitioners to help patients initiate and maintain a routine of daily exercise [113, 114].
An interesting finding from this pilot study was that eight weeks of exercise affected healthy and MDD brains similarly. The deactivation pattern we observed in several brain regions warrants further investigation with a larger sample size to allow a more robust statistical analysis. Future work must also include an MDD control group, as this will help us understand the magnitude of the effect of exercise, in combination with other therapies, on depressive symptomology and neural function. Moreover, there is a shortage of reporting sedentary behavior, physical activity levels, and cardiovascular fitness parameters in the MDD literature. As such, some of the differences observed in studies comparing MDD to controls might be confounded by a low-active lifestyle, which may be more prevalent in MDD. To address this gap, future research should compare “fit” and “low-fit” MDD groups to identify markers independent of cardiorespiratory fitness and unique to MDD. Furthermore, the exercise and cognitive literature have not established whether the psychological effects from engaging in exercise, independent of changes in fitness, are still beneficial to mental health and brain function. Future research must indicate whether the effects seen following an exercise intervention are associated with improved cardiovascular fitness or from the psychological benefits from engaging in exercise.
Lastly, MDD is a heterogeneous disorder, and it is likely to be a multifaceted interaction of psychological and neurobiological mechanisms that underlie or mediate the effects of exercise. Future research must consider using a combination approach of multimodal imaging techniques, behavioral assessments, and biochemical analysis to delineate the biological and clinical signatures of fit and unfit MDD populations. Once we are able to elucidate these key biomarkers unique to MDD, novel intervention strategies can then be designed to prevent or reverse neuropsychological pathologies such as MDD.
|Age (years)||37.25 (8.00)||8||20.63 (1.19)||8||14||<0.0001b|
|Body mass index (kg/m2)||28.33 (5.12)||8||28.29 (7.91)||8||14||0.993b|
|VO2max (ml.kg−1.min−1)||24.82 (8.00)||7||20.81 (6.48)||8||12||0.326b|
|BDI||41.75 (3.50)||8||5.88 (5.03)||8||14||<0.0001b|
|MoCA||24.63 (1.41)||8||26.13 (3.23)||8||14||0.248b|
Data are expressed as the mean with the standard deviation in parentheses. aPearson's chi-square. bStudent's t-test. VO2max: maximum oxygen consumption; BDI: Beck Depression Inventory; MoCA: Montreal Cognitive Assessment.
|n||Premean (SD)||Postmean (SD)||p||d||n||Premean (SD)||Postmean (SD)||p||d||F||p||dppc2|
|BDI||8||41.75 (3.50)||15.50 (10.43)||0.0004||2.89||8||5.88 (5.03)||3.25 (4.53)||0.072||0.545||30.42||<0.0001||5.34|
|MOCA||8||24.63 (1.41)||25.75 (2.38)||0.229||0.564||8||26.13 (3.23)||27.13 (1.81)||0.286||0.338||0.011||0.920||0.047|
|Memory task (high confidence (correct) %)||8||71.48 (9.50)||69.08 (12.03)||0.359||0.210||8||75.32 (9.29)||75.13 (9.07)||0.942||0.020||0.418||0.529||0.230|
|BMI (kg/m2)||8||28.33 (5.12)||28.29 (4.48)||0.934||0.007||8||28.29 (7.91)||27.95 (6.61)||0.530||0.023||0.205||0.657||0.044|
|VO2max (ml.kg−1.min−1)||7||24.82 (8.00)||32.52 (10.12)||0.073||0.834||8||20.81 (6.48)||30.57 (8.66)||0.014||1.26||7.96||0.661||0.279|
Data are expressed as mean with SD in parentheses. BDI: Beck Depression Inventory; MoCA: Montreal Cognitive Assessment; BMI: body mass index; VO2max: maximum oxygen consumption; d: Cohen's d.
|Voxels||Peak T||MNI coordinates||BA||Location|
|37||−6.27||−6||17||41||32||Medial superior frontal/mid cingulate|
|35||−5.79||−30||−46||44||40||Left supramarginal/intraparietal sulcus|
|45||−5.61||−42||−28||35||3||Postcentral, superior parietal gyrus|
|139||−5.3||−42||−37||8||41||Left posterior insula|
|30||−5.22||−18||−70||47||7||Left superior parietal|
|36||−5.17||30||−28||14||Right posterior insula|
|20||−4.75||54||−43||−4||21||Right posterior mid temporal|
|25||−3.82||12||−82||−7||18||Right occipital gyrus|
|Voxels||Peak T||MNI coordinates||BA||Location|
|27||−4.92||−39||−22||56||4||Left precentral gyrus|
Conflicts of Interest
The authors declare that there is no conflict of interest regarding the publication of this paper.
- 1. Cognitive functions in depressive disorders: evidence from a population-based studyPsychological Medicine20043418391[PubMed][Google Scholar]
- 2. Effect of antidepressant treatment on cognitive impairments associated with depression: a randomised longitudinal studyThe Lancet Psychiatry201635425435[PubMed][Google Scholar]
- 3. An analysis of memory dysfunction in major depressionJournal of the Affective Disorders19953519[Google Scholar]
- 4. Neuropsychologic impairments in bipolar and unipolar mood disorders on the CANTAB neurocognitive batteryBiological Psychiatry200048674685[PubMed][Google Scholar]
- 5. Putting names to faces: successful encoding of associative memories activates the anterior hippocampal formationNeuroImage200320214001410[PubMed][Google Scholar]
- 6. On the nature and pattern of neurocognitive function in major depressive disorderNeuropsychiatry, Neuropsychology, & Behavioral Neurology1998113111119[Google Scholar]
- 7. Episodic and declarative memory: role of the hippocampusHippocampus199883198204[PubMed][Google Scholar]
- 8. Depression and memory impairment: a meta-analysis of the association, its pattern, and specificityPsychological Bulletin19951172285305[PubMed][Google Scholar]
- 9. Hippocampal and brain stem activation during word retrieval after repeated and semantic encodingCerebral Cortex2002129p. 900[PubMed][Google Scholar]
- 10. Learning and memory: the hippocampus encodes objects in timeNature Reviews Neuroscience2014155p. 281[Google Scholar]
- 11. Episodic encoding and recognition of pictures and words: role of the human medial temporal lobesActa Psychologica20001052-3159179[PubMed][Google Scholar]
- 12. Impaired recognition memory in monkeys after damage limited to the hippocampal regionJournal of Neuroscience2000201p. 451[PubMed][Google Scholar]
- 13. Loss of recent memory after bilateral hippocampal lesionsJournal of Neurology, Neurosurgery, and Psychiatry19572011121[Google Scholar]
- 14. Hippocampal volume and depression: a meta-analysis of MRI studiesAmerican Journal of Psychiatry200416111p. 1957[Google Scholar]
- 15. A meta-analysis examining clinical predictors of hippocampal volume in patients with major depressive disorderJournal of Psychiatry & Neuroscience: JPN2009341p. 41[Google Scholar]
- 16. Subcortical brain alterations in major depressive disorder: findings from the ENIGMA major depressive disorder working groupMolecular Psychiatry2016216806812[PubMed][Google Scholar]
- 17. Memory related dysregulation of hippocampal function in major depressive disorderBiological Psychology2010853499503[PubMed][Google Scholar]
- 18. Abnormal hippocampal activation in patients with extensive history of major depression: an fMRI studyJournal of Psychiatry & Neuroscience: JPN2012371p. 28[Google Scholar]
- 19. Hippocampal activation during associative encoding of word pairs and its relation to symptomatic improvement in depression: a functional and volumetric MRI studyJournal of Affective Disorders2014152–154462467[PubMed][Google Scholar]
- 20. Gliogenesis and glial pathology in depressionCNS & Neurological Disorders Drug Targets200763219233[PubMed][Google Scholar]
- 21. The neuropathology of primary mood disorderBrain2002125714281449[PubMed][Google Scholar]
- 22. Stress, depression, and neuroplasticity: a convergence of mechanismsNeuropsychopharmacology200733188109[PubMed][Google Scholar]
- 23. Cellular changes in the postmortem hippocampus in major depressionBiological Psychiatry2004569640650[PubMed][Google Scholar]
- 24. Progressive alterations of hippocampal CA3-CA1 synapses in an animal model of depressionBehavioural Brain Research2014275191200[PubMed][Google Scholar]
- 25. Using epidemiological data to model efficiency in reducing the burden of depressionThe Journal of Mental Health Policy and Economics200034175186[PubMed][Google Scholar]
- 26. Long-term treatment of depression: there is more than drugsRecenti Progressi in Medicina2002936p. 343[Google Scholar]
- 27. Depression: an undertreated disorderHealth Reports199784918[Google Scholar]
- 28. Research on major depression: strategies and prioritiesThe Journal of the American Medical Association200328923p. 3167[Google Scholar]
- 29. Exercise treatment for major depression: maintenance of therapeutic benefit at 10 monthsPsychosomatic Medicine20006221633638[PubMed][Google Scholar]
- 30. Effects of exercise training on older patients with major depressionArchives of Internal Medicine19991591923492356[PubMed][Google Scholar]
- 31. Exercise and pharmacotherapy in the treatment of major depressive disorderPsychosomatic Medicine2007697p. 587[PubMed][Google Scholar]
- 32. Effects of aerobic exercise on depression: a controlled studyBritish Medical Journal19852916488p. 109[Google Scholar]
- 33. Exercise interventions for mental health: a quantitative and qualitative reviewClinical Psychology: Science and Practice2006132179193[Google Scholar]
- 34. Exercise as a treatment for depression: a meta-analysis adjusting for publication biasJournal of Psychiatric Research2016774251[PubMed][Google Scholar]
- 35. Exercise: a behavioral intervention to enhance brain health and plasticityTrends in Neurosciences2002256295301[PubMed][Google Scholar]
- 36. Running increases cell proliferation and neurogenesis in the adult mouse dentate gyrusNature Neuroscience199923266270[PubMed][Google Scholar]
- 37. All about running: synaptic plasticity, growth factors and adult hippocampal neurogenesisCurrent Topics in Behavioral Neurosciences201315189210[PubMed][Google Scholar]
- 38. Exercise enhances learning and hippocampal neurogenesis in aged miceJournal of Neuroscience20052538[Google Scholar]
- 39. Neurogenesis and exercise: past and future directionsNeuromolecular Medicine2008102128140[PubMed][Google Scholar]
- 40. Running is the neurogenic and neurotrophic stimulus in environmental enrichmentLearning & Memory (Cold Spring Harbor, N.Y.)2011189605609[PubMed][Google Scholar]
- 41. Exercise-induced synaptogenesis in the hippocampus is dependent on UCP2-regulated mitochondrial adaptationJournal of Neuroscience20082842[Google Scholar]
- 42. Insulin-like growth factor I is required for vessel remodeling in the adult brainProceedings of the National Academy of Sciences of the United States of America20041012698339838[PubMed][Google Scholar]
- 43. Exercise-induced changes in dendritic structure and complexity in the adult hippocampal dentate gyrusNeuroscience2006137412991307[PubMed][Google Scholar]
- 44. Hippocampal neurogenesis and dendritic plasticity support running-improved spatial learning and depression-like behaviour in stressed ratsPLoS One201169, article e24263[PubMed][Google Scholar]
- 45. Exercise training increases size of hippocampus and improves memoryProceedings of the National Academy of Sciences of the United States of America2011108730173022[PubMed][Google Scholar]
- 46. Exergaming and older adult cognition: a cluster randomized clinical trialAmerican Journal of Preventive Medicine2012422109119[PubMed][Google Scholar]
- 47. Plasticity of brain networks in a randomized intervention trial of exercise training in older adultsFrontiers in Aging Neuroscience20102p. 32[PubMed][Google Scholar]
- 48. The influence of aerobic fitness on cerebral white matter integrity and cognitive function in older adults: results of a one-year exercise interventionHuman Brain Mapping2013341129722985[PubMed][Google Scholar]
- 49. Aerobic exercise to improve cognitive function in older people without known cognitive impairmentCochrane Database of Systematic Reviews2015John Wiley & Sons, LtdChichester, UK[Google Scholar]
- 50. The effect of exercise training on cognitive function in older adults with mild cognitive impairment: a meta-analysis of randomized controlled trialsThe American Journal of Geriatric Psychiatry2013211110861097[PubMed][Google Scholar]
- 51. Aerobic exercise and neurocognitive performance: a meta-analytic review of randomized controlled trialsPsychosomatic Medicine2010723239252[PubMed][Google Scholar]
- 52. Effect of a 24-month physical activity intervention vs health education on cognitive outcomes in sedentary older adultsThe Journal of the American Medical Association20153148p. 781[PubMed][Google Scholar]
- 53. Aerobic exercise training increases brain volume in aging humansThe Journals of Gerontology Series A: Biological Sciences and Medical Sciences20066111p. 1166[PubMed][Google Scholar]
- 54. Aerobic fitness is associated with hippocampal volume in elderly humansHippocampus2009191010301039[PubMed][Google Scholar]
- 55. Hippocampal plasticity in response to exercise in schizophreniaArchives of General Psychiatry2010672p. 133[PubMed][Google Scholar]
- 56. American Psychiatric Association, and American Psychiatric AssociationTask Force on DSM-IV, Diagnostic and Statistical Manual of Mental Disorders: DSM-IV-TR2000Washington, DCAmerican Psychiatric Publishing, Inc.
- 57. BDI-II. Beck Depression Inventory Second Edition. Manual1996San AntonioPsychological Corporation
- 58. The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairmentJournal of the American Geriatrics Society2005534695699[PubMed][Google Scholar]
- 59. Comparison of Beck depression inventories-IA and-II in psychiatric outpatientsJournal of Personality Assessment1996673588597[PubMed][Google Scholar]
- 60. Cross-validation of the YMCA submaximal cycle ergometer test to predict VO2maxResearch Quarterly for Exercise and Sport2004753337342[PubMed][Google Scholar]
- 61. Y’s way to Physical Fitness: The Complete Guide to Fitness Testing and Instruction19893rdChampaign, ILPublished for YMCA of the USA by Human Kinetics Publishers106108
- 62. ACSM’s Guidelines for Exercise Testing and Prescription20149PhiladelphiaWolters Kluwer/Lippincott Williams & Wilkins Health
- 63. Canada’s physical activity guide to healthy active living for older adultsJournal of Aging and Physical Activity2004123Champaign, USAHuman Kinetics Publ INC.[Google Scholar]
- 64. American College of Sports Medicine position stand. Quantity and quality of exercise for developing and maintaining cardiorespiratory, musculoskeletal, and neuromotor fitness in apparently healthy adults: guidance for prescribing exerciseMedicine and Science in Sports and Exercise201143713341359[PubMed][Google Scholar]
- 65. Exercise and the treatment of depression: a review of the exercise program variablesJournal of Science and Medicine in Sport2014172177182[PubMed][Google Scholar]
- 66. Exercise for depressionJournal of Evidence-Based Medicine201364307308[Google Scholar]
- 67. Estimating effect sizes from the pretest-posttest-control group designsOrganizational Research Methods2008112364386[Google Scholar]
- 68. A mini-review of fMRI studies of human medial temporal lobe activity associated with recognition memoryQuarterly Journal of Experimental Psychology B2005583-4340360[PubMed][Google Scholar]
- 69. The dorsolateral prefrontal cortex contributes to successful relational memory encodingJournal of Neuroscience20072720[Google Scholar]
- 70. Dorsolateral prefrontal cortex promotes long-term memory formation through its role in working memory organizationJournal of Neuroscience2006263[Google Scholar]
- 71. Neural correlates of memory for items and for associations: an event-related functional magnetic resonance imaging studyJournal of Cognitive Neuroscience2005174652667[PubMed][Google Scholar]
- 72. Relational memory and the hippocampus: representations and methodsFrontiers in Neuroscience200932166174[PubMed][Google Scholar]
- 73. Observing the transformation of experience into memoryTrends in Cognitive Sciences20026293102[PubMed][Google Scholar]
- 74. Building memories: remembering and forgetting of verbal experiences as predicted by brain activityScience19982815380[PubMed][Google Scholar]
- 75. Exercise as an add-on strategy for the treatment of major depressive disorder: a systematic reviewCNS Spectrums2014196496508[PubMed][Google Scholar]
- 76. Cognitive neuroscience: forgetting of things pastCurrent Biology20011123R964R967[PubMed][Google Scholar]
- 77. Cognitive neuroscience: forgetting of things pastCurrent Biology20011123R964R967[PubMed][Google Scholar]
- 78. Age effects on the neural correlates of successful memory encodingBrain20031261213229[PubMed][Google Scholar]
- 79. Prefrontal-temporal circuitry for episodic encoding and subsequent memoryJournal of Neuroscience20002016[Google Scholar]
- 80. When more means less: neural activity related to unsuccessful memory encodingCurrent Biology2001111915281530[PubMed][Google Scholar]
- 81. When less means more: deactivations during encoding that predict subsequent memoryNeuroImage2004233921927[PubMed][Google Scholar]
- 82. Age-related differences in neural activity during memory encoding and retrieval: a positron emission tomography studyThe Journal of Neuroscience1997171391400[PubMed][Google Scholar]
- 83. Role of human prefrontal cortex in attention controlAdvances in Neurology1995662134[PubMed][Google Scholar]
- 84. Paying attention to the thalamic reticular nucleusTrends in Neurosciences19982112832[PubMed][Google Scholar]
- 85. GABAergic contributions to the pathophysiology of depression and the mechanism of antidepressant actionCNS & Neurological Disorders-Drug Targets200762127140[PubMed][Google Scholar]
- 86. The GABAergic deficit hypothesis of major depressive disorderMolecular Psychiatry2011164383406[PubMed][Google Scholar]
- 87. Don’t worry ‘B’ happy!: a role for GABAB receptors in anxiety and depressionTrends in Pharmacological Sciences20052613643[PubMed][Google Scholar]
- 88. Evidence for impaired cortical inhibition in patients with unipolar major depressionBiological Psychiatry2006595395400[PubMed][Google Scholar]
- 89. Reduction in occipital cortex γ-aminobutyric acid concentrations in medication-free recovered unipolar depressed and bipolar subjectsBiological Psychiatry2007616806812[PubMed][Google Scholar]
- 90. Increased cortical GABA concentrations in depressed patients receiving ECTThe American Journal of Psychiatry20031603[PubMed][Google Scholar]
- 91. Reduced prefrontal glutamate/glutamine and γ-aminobutyric acid levels in major depression determined using proton magnetic resonance spectroscopyArchives of General Psychiatry2007642p. 193[Google Scholar]
- 92. Reduced density of calbindin immunoreactive GABAergic neurons in the occipital cortex in major depression: relevance to neuroimaging studiesBiological Psychiatry2010675465470[PubMed][Google Scholar]
- 93. GABAergic neurons immunoreactive for calcium binding proteins are reduced in the prefrontal cortex in major depressionNeuropsychopharmacology2007322471482[PubMed][Google Scholar]
- 94. Physical exercise prevents stress-induced activation of granule neurons and enhances local inhibitory mechanisms in the dentate gyrusThe Journal of Neuroscience2013331877707777[PubMed][Google Scholar]
- 95. The effect of exercise training in improving motor performance and corticomotor excitability in people with early Parkinson’s diseaseArchives of Physical Medicine and Rehabilitation200889712211229[PubMed][Google Scholar]
- 96. An 8-month randomized controlled exercise trial alters brain activation during cognitive tasks in overweight childrenObesity2014221232242[PubMed][Google Scholar]
- 97. A 12-week physical and cognitive exercise program can improve cognitive function and neural efficiency in community-dwelling older adults: a randomized controlled trialJournal of the American Geriatrics Society201563713551363[PubMed][Google Scholar]
- 98. Semantic memory functional MRI and cognitive function after exercise intervention in mild cognitive impairmentJournal of Alzheimer’s Disease2013371197215[PubMed][Google Scholar]
- 99. Differences in brain activity during a verbal associative memory encoding task in high- and low-fit adolescentsJournal of Cognitive Neuroscience2013254595612[PubMed][Google Scholar]
- 100. Neuroplasticity—exercise-induced response of peripheral brain-derived neurotrophic factorSports Medicine2010409765801[PubMed][Google Scholar]
- 101. Increased neurogenesis and the ectopic granule cells after intrahippocampal BDNF infusion in adult ratsExperimental Neurology20051922348356[PubMed][Google Scholar]
- 102. Brain-derived neurotrophic factor (bdnf) can prevent apoptosis of rat cerebellar granule neurons in cultureDevelopmental Brain Research1995852249258[PubMed][Google Scholar]
- 103. The BDNF effects on dendritic spines of mature hippocampal neurons depend on neuronal activityFrontiers in Synaptic Neuroscience20146p. 5[PubMed][Google Scholar]
- 104. BDNF signaling in the formation, maturation and plasticity of glutamatergic and GABAergic synapsesExperimental Brain Research20091993-4203234[PubMed][Google Scholar]
- 105. BDNF promotes differentiation and maturation of adult-born neurons through GABAergic transmissionThe Journal of Neuroscience201232411431814330[PubMed][Google Scholar]
- 106. A biological function for the neuronal activity-dependent component of Bdnf transcription in the development of cortical inhibitionNeuron2008604610624[PubMed][Google Scholar]
- 107. Impairment of the spatial learning and memory induced by learned helplessness and chronic mild stressPharmacology Biochemistry and Behavior2006832186193[PubMed][Google Scholar]
- 108. Brain-derived neurotrophic factor produces antidepressant effects in behavioral models of depressionJournal of Neuroscience2002228p. 3251[PubMed][Google Scholar]
- 109. Exercise-induced regulation of brain-derived neurotrophic factor (BDNF) transcripts in the rat hippocampusMolecular Brain Research1998611147153[PubMed][Google Scholar]
- 110. Exercise builds brain health: key roles of growth factor cascades and inflammationTrends in Neurosciences2007309464472[PubMed][Google Scholar]
- 111. Voluntary exercise produces antidepressant and anxiolytic behavioral effects in miceBrain Research20081199148158[PubMed][Google Scholar]
- 112. Too little exercise and too much sitting: inactivity physiology and the need for new recommendations on sedentary behaviorCurrent Cardiovascular Risk Reports200824292298[PubMed][Google Scholar]
- 113. Depression is a risk factor for noncompliance with medical treatment: meta-analysis of the effects of anxiety and depression on patient adherenceArchives of Internal Medicine20001601421012107[PubMed][Google Scholar]
- 114. Opinion and evidence: is exercise a viable treatment for depression?ACSM’s Health and Fitness Journal20121641421[PubMed][Google Scholar]