The inevitable effects of ageing on the physical and mental health of humans can somewhat be offset by regular participation in physical exercise during adulthood [1]. Although an acute bout of exercise positively affects executive cognitive performance, regular exercise during midlife has a protective effect against cognitive decline in the later adult years [2].

Etnier & Chang [3] described executive function as “a higher order cognitive ability that controls basic, underlying cognitive functions for purposeful, goal-directed behaviour and that has been associated with frontal lobe activity.” Executive function refers to the domains of cognitive function that involve executive control, including planning, scheduling, working memory, interference control and task coordination [4]. Conceptually, executive function is considered critical for performance in novel situations or when an individual is required to inhibit a previously learned response [3].

McAuley et al. [5] asserted that a decline in executive cognitive control is associated with the normal ageing process. This proposed decline has been associated with changes (e.g. volumetric) in certain brain areas, especially the frontal lobes. Additionally, Royall et al. [6] demonstrated in a three-year cohort study that the regression in executive cognitive control is independently associated with longitudinal declines in functional status.

Tests designed to assess executive function usually represent external tasks that are unfamiliar or uncommon and requires an individual to apply certain intellectual abilities (e.g. planning) in order to solve the task/problem [7]. The Stroop task [8] is one of the most frequently used executive function measures. A number of modifications of this test (i.e. card-based and computerized) have been used in research and clinical settings. The Stroop task, consisting of conditions of increasing difficulty, is used to assess a number of executive function components, including selective attention, the ability to shift response/perceptual sets and the ability to inhibit habitual responses [9].

Recently, there has been growing interest in the promotion of physical activity to improve cognitive function and there is mounting evidence that exercise can positively influence and preserve this construct. Furthermore, it was suggested that the greatest effects are observed for higher level cognitive functions, such as executive function, with fewer effects on lower level functions [10]. Up till now the majority of longitudinal studies focussed on the effect of a single exercise training modality on executive cognitive function, i.e. aerobic exercise [11–13] or resistance exercise [14–17]. Studies comparing the effects of aerobic and resistance training on cognitive function are in the minority. Nevertheless, from what is known so far, it would seem that aerobic training yields the best results [18, 19].

The majority of longitudinal resistance training studies report improvements in executive cognitive processes [14–17, 20, 21]. It was also suggested that aerobic training improves performance on tasks which demand greater executive control processes, a phenomenon known as the “selective improvement” hypothesis [18, 22].

Aerobic training interventions generally have a more profound impact on cardiovascular fitness compared to resistance training [12, 19]. However, high-intensity aerobic interval training (HIIT) has been shown to induce larger increases in maximal aerobic capacity (VO_2max) compared to aerobic training at a constant intensity [23, 24]. Consequently, as cardiovascular fitness has been proposed as a potential mediating factor in the enhancement of cognitive performance [13], the question regarding the effect of HIIT on cognition is a matter of interest.

Therefore, the present study aimed to determine if different exercise training modalities (resistance training, high-intensity aerobic interval training and moderate continuous aerobic training) have similar effects on the cognitive performance of older individuals, as assessed by a Stroop task. Furthermore, the effects of the different training interventions on measures of physical function, i.e. walking endurance and functional mobility, were investigated. The authors’ primary hypothesis was that high intensity interval aerobic training will have superior effects on the physical and cognitive function of older adults compared to resistance training and moderate continuous aerobic training. The secondary hypothesis was that the exercise training groups will show greater improvements on the executive function tasks compared to the information processing speed task.

To the best of our knowledge, this is the first attempt to compare the effects of high-intensity aerobic interval training to traditional (continuous) aerobic training, as well as resistance training on cognitive function in older adults. Moderate continuous aerobic training (MCT) proved to be most beneficial for the enhancement of executive cognitive function, a higher level cognitive process; whereas high-intensity aerobic interval training (HIIT) had the greatest positive effect on information processing speed, a lower level cognitive process. Resistance training (RT) at a moderate intensity was more beneficial for gains in information processing speed compared to MCT, and executive cognitive function compared to HIIT. As hypothesized, the greatest improvement in physical function, i.e. walking endurance and functional mobility, was induced by HIIT. However, HIIT was the only study group that did not show a statistically significant improvement in performance on the executive function tasks, a finding contradictory to the authors’ hypothesis.

Accuracy on the two Stroop tasks did not change significantly after the 16-week period in either of the study groups. A moderate decrease in accuracy on the Incongruent task was, however, observed in the control (CON) group after the intervention period. It is therefore suggested that the CON group’s improvement in reaction time on the Incongruent Stroop task could be due to a speed-accuracy trade off.

The cognitive function results of the present study, specifically the changes exhibited by the MCT group, provide support for the “selective improvement” hypothesis, as proposed by Kramer et al. [18]. These authors also observed a selective effect of aerobic exercise on cognitive function. Thus, other performance measures that were not linked to executive function remained unaffected. Furthermore, our results are in agreement with previous intervention studies which proposed that different types of exercise interventions have unique effects on cognition [31]. In addition, our results extend the existing literature by adding novel findings with regards to the effect of HIIT on cognitive function.

Traditional moderate-intensity aerobic training showed the greatest benefit on tasks assessing executive function, with no significant improvements in information processing speed. The findings of Predovan et al. [13] also provide a degree of support for the hypothesis that traditional aerobic exercise training has a selective effect on cognition. After a 12-week intervention, the training group improved their performance in the inhibition/switching (i.e. set shifting) task (Stroop Interference), which was considered to recruit the highest level of executive functioning (multiple executive processes), but no training effect was found for the naming (Neutral) or inhibition (Incongruent) conditions of the Stroop task. They therefore suggested that aerobic dancing selectively improves tasks assessing switching and not necessarily tasks requiring inhibitory processes. The researchers proposed that differences in the type of aerobic exercises performed (i.e. aerobic dancing, walking, running etc.) could explain the inconsistent findings across studies [13].

In contrast to the findings of the present study, some investigators observed a beneficial effect of aerobic training at a constant, moderate intensity on tasks assessing information processing speed (lower level cognitive function) in addition to executive function. A meta-analytic review by Smith et al. [32] reported modest improvements in older adults’ lower and higher level neurocognitive functions after participation in aerobic training interventions that lasted between 6 weeks and 18 months. These improvements were observed for attention and processing speed, executive function and memory; whereas working memory did not benefit from aerobic training [32]. Additionally, Dustman et al. [12] found that 4 months of aerobic training resulted in significantly better scores on simple and complex cognitive tasks compared to strength and flexibility training.

In the present study, aerobic interval training at alternating periods of high and low intensities significantly improved performance on a task assessing information processing speed, whereas no improvement was found on the executive function tasks. Thus, both the HIIT and MCT interventions induced selective improvements in cognitive function, albeit different cognitive domains were affected.

In a more recent study, conducted by Iuliano et al. [31], 12 weeks of cardiovascular training at higher intensities significantly improved attention and abstract reasoning; whereas no improvement was observed in older adults’ executive function. The dissimilarities in the study outcomes could be attributed to the differences in training stimuli (i.e. shorter duration training sessions at a higher intensity in the former study vs longer duration training sessions at a moderate intensity in the present study).

In the present study RT at a moderate intensity was more beneficial for gains in information processing speed compared to MCT, and executive cognitive function compared to HIIT. These findings do not support the notion that RT has a selective effect on cognition, as positive results were obtained across all domains (higher and lower level) of cognitive function measured. The majority of previous intervention studies reported a selective improvement of RT on cognitive function, with executive control tasks showing the greatest improvement [14–17, 20, 21]. Additionally, positive associations have been demonstrated between greater lower extremity strength and better executive function [33], as well as muscle-strengthening activities and executive function [34]. Researchers observed an improvement in the Stroop Incongruent task in older adults after 4 weeks of RT, however, the same effect was not found on a task assessing information processing speed [14], amounting to a selective effect on cognitive function. This finding is contradictory to the positive effect of RT found for both executive function and information processing speed in the present study. One could argue that the study duration of 4 weeks chosen by Anderson-Hanley et al. [14] was too short to induce noteworthy improvements in information processing speed, but that longer duration RT can have beneficial effects. Liu-Ambrose et al. [16] reported an improvement in the Stroop Interference task after 12 months of RT, compared to balance and tone exercises in community-dwelling older women. Smiley-Oyen et al. [19] found an improvement in the Stroop Incongruent task after 10 months of aerobic training in older men and women, but in contrast to the results of the present study, no effect was observed in the strength-and-flexibility training group. This group performed exercises with resistance bands, free weights and stability balls. It could be argued that the intensity of the strength exercises was too low to provide a sufficient stimulus for cognitive improvements. Furthermore, their results could only partly corroborate the “selective improvement” hypothesis, as positive results were not obtained across all the tasks assessing executive function. The researchers neglected to include a no-exercise control group, leading one to question whether the results can be solely attributed to the exercise.

As hypothesized, the HIIT group experienced the largest improvement in walking endurance compared to the other training groups. This is not an uncommon finding in the literature [23, 24]. Helgerud et al. [23] found that healthy, trained men who exercised at higher intensities (90–95% HR_max) exhibited the biggest gains in aerobic capacity, whereas those exercising at lower intensities (70 and 85% HR_max) did not improve their VO_2max. This finding, as well as the results of the current study, supports the notion that cardiovascular adaptations to training (i.e. improvements in VO_2max) are intensity dependent [35].

A surprising finding was that the MCT group did not experience an improvement in walking endurance, while the RT group actually showed a greater practically significant improvement in walking endurance compared to the MCT group. It is generally accepted that aerobic training leads to larger increases in aerobic capacity compared to resistance training [12, 19]. However, while some investigators observed no effect of RT on VO_2max [20], others reported a beneficial effect on VO_2max and walking endurance [36–38].

The results exhibited by the MCT group are also inconsistent with the findings reported by Dustman et al. [12], where a pronounced increase in VO_2max and Stroop Interference was observed after 16 weeks of aerobic training. There may be two possible explanations for these conflicting findings: (a) differences in computing the Stroop Interference effect and (b) dissimilarities in the exercise protocols. Dustman et al. [12] subtracted the reaction time for the Stroop Word task from the Incongruent task. However, the Interference score in the present study was calculated as the difference in reaction time between the Neutral and Incongruent tasks, similar to a more recent intervention study [16]. Furthermore, the intensity reported in Dustman’s study was higher than the exercise intensity used in the present study. It could thus be argued that our training stimulus was too low for the MCT group to induce significant improvements in aerobic capacity.

Mechanisms such as angiogenesis, synaptogenesis and neurogenesis have been proposed as possible mediating factors in the exercise-cognition relationship. Increases in biological mediators, including brain-derived neurotrophic factor (BDNF) and insulin-like growth factor 1 (IGF-1), have also been linked to exercise training [15, 39]. However, it has been suggested that the upregulation of these biological mediators are not dependent on aerobic fitness [19] and that neurocognitive networks are differentially influenced by the exercise training mode [40]. Aerobic training is linked to elevated levels of BDNF [39], while resistance training produces increased levels of IGF-1 [15]. These dissimilarities could possibly serve as an explanation for the differential effects of exercise training mode on cognitive function observed in the present study.

Positive effects of aerobic training on cognitive function and cerebral oxygenation during cortical activation have been recently reported in the literature [41]. We previously demonstrated that changes in cerebral oxygenation are differentially influenced by the exercise training mode [42] and we therefore suggest that the MCT group, irrespective of their minimal changes in aerobic fitness, experienced the most profound structural and functional neural adaptations which enabled them to perform practically significantly better, i.e. more efficiently, on the executive function tasks compared to the other study groups.

MCT and RT proved to be superior to HIIT for the enhancement of older individuals’ executive cognitive function; whereas HIIT were most beneficial for the improvement in information processing speed. HIIT also induced the largest gains in physical function, i.e. walking endurance and functional mobility.