Emerging research suggests that exercise has beneficial effects on brain health, and its role as a therapeutic is under investigation in various neurological conditions. In order to determine the optimal dose-response of exercise for different patient populations, the dynamic mechanisms by which exercise exerts affects the brain must be understood.

Whereas previous work has investigated exercise effects on the brain in isolation¹, in this work we consider a range of exercise-induced systemic effects which may determine cerebrovascular changes. Thus, we examine group-level effects of a single session of aerobic exercise along with the interaction between systemic and cerebral physiology after exercise using multi-modal fMRI techniques.

22 healthy participants (11 males, 26.6 ± 4.7 years old) had an MRI scan on a 3T GE HDx system. The baseline MRI session included a multi-TI (mTI) arterial spin labelling (ASL) sequence (TIs: 300,400,500, 600,700,800, 1100, 1400, 1700, 2000ms, TE = 0.27ms; slice delay = 52 ms, QUIPSS II² cutoff at TI>700 ms) to measure cerebral blood flow (CBF), a hypercapnia gas challenge (targeted +5 mmHg P_ETCO₂) with a dual-echo gradient echo spiral readout sequence (TR = 2.2s; TE₁ = 2.7s; TE₂ =29ms; 64 x 64; 15 slices; resolution = 3.1 × 3.1 × 8.4 mm³) to measure cerebrovascular reactivity (CVR) in a subset of participants (n=12), and a structural FSPGR (1mm³ resolution) sequence.

Participants then completed 20-minutes of moderate intensity aerobic exercise on a cycle ergometer. Immediately after, participants had a repeat MRI scan with the multi-TI sequence repeated 3 times (see Fig.1) and the hypercapnia dual-echo sequence repeated at post 30-minutes. Perfusion quantification was performed on a voxel-by-voxel basis using a two-compartment model³. For the mTI ASL, physiological noise (RETROICOR⁴, P_ETCO₂, respiration volume (RVT) and heart rate (HR)) were regressed out. For BOLD and CBF CVR, following surround averaging and subtraction respectively, P_ETCO₂ was used as a regressor in a general linear model⁵.

Exercise-related physiological parameters of interest were: [1] HR recovery (difference between HR at peak exercise and 1 minute after cessation), with attenuated HR recovery a presumed index of reduced parasympathetic activity⁶, [2] average blood lactate concentration during exercise, as a measure of exercise-induced physiological strain, [3] mean arterial pressure following exercise.

Age and body mass index were found to significantly correlate with CBF post exercise, and were added as covariates in the CBF mTI analysis. A repeated-measures ANOVA found a significant main effect of time on GM CBF after accounting for age and BMI (F _3,54 = 2.85, p < 0.05); however post-hoc analyses showed this was driven by differences between post-exercise sessions which did not survive multiple comparison correction. Neither BOLD CVR nor CBF CVR were significantly altered by exercise when measured 30-minutes post exercise cessation (Fig.2, CBF CVR p = 0.072).

Despite the lack of an effect of exercise on cerebrovascular measures on a group level, the exercise intervention resulted in prolonged systemic physiological changes which endured for up to 40-minutes post exercise (Fig.3) which were significantly associated with the MR cerebrovascular measures (see Table 1). Average blood lactate concentration (mmol/L) measured during the exercise intervention significantly predicted the change in CBF 40-minutes after exercise (Fig.2D); a higher lactate concentration, indicative of higher physiological load, was associated with a reduction in CBF after exercise, whereas a lower lactate concentration was associated with increased CBF. Mean arterial pressure predicted CBF 20-minutes post exercise, whilst HR recovery from exercise was associated with baseline CVR (CBF CVR r= -0.609; BOLD CVR r = -0.595, both p < 0.05 uncorrected) and change in CBF post-20minutes, with a quicker recovery associated with less change.

Despite enduring systemic physiological recovery in the 1-hour period following exercise cessation, absolute CBF and hypercapnia-induced CVR were not significantly changed from baseline when assessed at the group level. Crucially, this demonstrates robust cerebral autoregulation in a healthy young population and contradicts previous smaller studies^1,7 showing altered cerebrovasculature post exercise without accounting for systemic physiology. The lack of a main effect of exercise on cerebrovascular MR measures may suggest that unlike cardiovascular health, the cerebrovascular health benefits associated with exercise are driven purely by chronic adaptive mechanisms rather than a combination of both acute and chronic mechanisms in a healthy population. Notably, the change in CBF after exercise was sensitive to systemic physiological factors related to exercise intensity, with the change in CBF found to be dependent on a physiological index of metabolic load.

This understanding, of systemic physiological predictors of the cerebrovascular response to exercise, will ultimately bring us closer to individually-tailored exercise prescription.