Background
Age- and behavior-related changes in the human brain involve structural, functional, and metabolic levels. Age- and behavior-associated alterations in white matter integrity, grey matter volume [
1‐
3], and neurotransmitter (e.g. dopamine, serotonin, and acetylcholine) synthesis and binding [
4‐
8] go along with deteriorations of cognitive functions, e.g. executive functions (EFs). EFs are higher-level cognitive functions that control and guide lower-level cognitive functions and goal-directed actions [
9], such as walking in challenging environments. Gait performance is partially controlled by different EF components, e.g. “working memory” [
10], “inhibition” [
11], and “divided attention” [
12]. Especially divided attention is associated with temporal and spatial dual-task gait parameters [
13]. Gait disturbances and falls seem to be related to the quality of EFs [
14,
15].
The prefrontal cortex (PFC), especially the dorsolateral prefrontal cortex and connected brain structures, has been linked with EFs [
16,
17]. Better EFs are associated with a greater PFC thickness and a larger PFC volume [
18]. During life, the (pre)frontal structure undergoes transformation processes; however, no agreement exists on the specific pattern of EFs adaptations [
1,
17,
19,
20]. One presumption is that a decrease in frontal grey matter volume and white matter integrity might be related to a decline of EFs. Rosano et al. (2012) illustrated that a smaller PFC volume might contribute to slower gait performance due to decreased information processing capacity [
21]. Furthermore, disrupted communication of cortico-cortical and cortico-subcortical networks, e.g. connection of frontal parts with parietal lobe and basal ganglia, respectively, are common causes of higher-level gait disorders [
15,
22]. Consequently, strengthening of EFs might improve gait performance [
23] and concomitantly might reduce falls as the risk of future falls can be predicted by EFs performance in older adults [
24].
Up to now, training of specific cognitive functions (e.g. EFs) may represent a central method to support specific brain functions and also preserve mobility in older adults [
21,
25‐
27]. Nonetheless, recent reviews examining the interaction of cognitive and physical functions concluded that a combined motor-cognitive training seems to be important for clinical practice to achieve safe movements in daily environment [
27‐
30]. On neuronal level, physical training triggers brain plasticity by cell proliferation and synaptic plasticity, while cognitive training seems to support the survival of newborn neurons and their integration in pre-existing networks [
28,
31,
32]. Especially, computerized training interventions seem effective [
27,
28,
33] when providing training principles that support (motor) learning [
33]. Video game-based physical exercises, or so-called exergames, allow concurrent training of motor and cognitive abilities. Incorporated video games are promising to train various cognitive functions [
34,
35]. Physical exercise (PE) interventions with decision-making opportunities are potentially able to improve both motor performance and cognition [
36]. Recent studies showed positive effects of exergame training on EFs and gait performance under dual-task condition in older adults [
37,
38] and a meta-analysis revealed that both healthy older adults and clinical populations with conditions associated with neurocognitive impairments benefit from physical-active video games [
39].
In various review articles, it is hypothesised that the impact of PE on the brain can be supported by concurrent intake of specific nutrients [
40‐
44]. This would mean, as a way of example, that a combination of PE with a nutritional supplement (NS) might further intensify the effects of PE on brain structure and function in older adults. The possible interplay between PE and nutrition involves common cellular processes essential for synaptic plasticity, neurogenesis, cell survival, and vascular function [
40‐
44].
Nonetheless, a recent systematic review concluded that former studies executing a combined approach of PE and NS to evoke neuronal adaptations were not particularly successful due to the misfit between the combinations; the elements were not chosen based on sharing of similar neuronal mechnism [
45]. The review argues, however, that especially omega-3 fatty acids (FAs), present in fish oil, might be an efficient NS promoting the beneficial effects of PE. Omega-3 FAs are essential for energy metabolism, for the function and integrity of the neuronal plasma membranes (with docosahexaenoic acid (DHA), arachidonic acid, and eicosapentaenoic acid (EPA) as their main components), and for blood perfusion in the brain [
46,
47]. Particularly, older adults may profit from FA supplementation, as in the aging brain the concentration of long chain polyunsaturated FAs (LCPUFAs) concentration decreases [
46]. LCPUFAs intake improves cognition, decreases (neuro)inflammation, and reduces vascular risk factors in older adults [
46]. On brain level, LCPUFAs may have positive effects on neuronal structure, function, and cerebral blood flow [
48]. For example, DHA acts as a neurotrophic factor by increasing the level of the brain-derived neurotrophic factor [
49]. Previous randomized-controlled studies showed that fish oil enhanced brain structure and function in healthy older adults, and participants improved working memory, EFs, white matter microstructure integrity, grey matter volume, and vascular parameters [
50,
51].
So far, studies could show that DHA supplementation enhanced the effects of exercise on axonal growth, brain derived neurotrophic factor-related synaptic plasticity, and cognition in rats [
49,
52]. However, no study exists that examined the combined effect of exergame training and omega-3 FAs on the brain in healthy older adults. This study, therefore, aims to investigate whether the positive effects of exergame training can be enhanced through adding omega-3 FA supplementation. The following research question guided the research process: “Does the combination of exergame training and fish oil differently affect neuronal system levels in the elderly brain compared to exergame training alone?” The main objectives of this study were to determine the effects of the intervention on corticospinal excitability and neuronal activity. We hypothesized that the combination would differently affect these parameters.
Discussion
The aim of this study was to investigate whether the known positive effects of exergame training can be enhanced by adding omega-3 FA supplementation. We hypothesized that the combination of exergame training and omega-3 FAs would differently affect neuronal system levels in the elderly brain compared to exergame training alone. Based on previous studies, we assumed that exergame training has positive effects on the elderly brain [
37,
38]. Furthermore, previous studies showed that omega-3 FAs have positive effects on the elderly brain [
85]. Although our results confirmed previous findings [
27,
37‐
39] by showing overall improvements in some of the outcomes (time main effects), the results showed no significant time × group interaction effects in any of the primary and secondary parameters. For the blood values, significant time × group interaction effects were measured. The fish oil intake group showed a significant increase of the omega-3 FAs. This increase indicates that the participants adhered to their intake schedule which led to the increased omega-3 FA levels within the first 16 weeks. However, this increase did not lead to any additional benefits in chosen outcomes due to adding fish oil to the exergame intervention. One reason might be that the exergame training acted as the main factor evoking effects while omega-3 FAs played a subordinate role. This assumption would be in line with a recently published pilot study where the combined approach of aerobic exercise and cognitive stimulation with omega-3 FAs showed an effect on gray matter volume and sole omega-3 FAs intake in combination with placebo exercise in form of stretching and toning did not induce effects [
86]. Another reason might be that an interplay exists but that our intervention study was not able to evoke and capture the effects. Several reasons may be given for this explanation. The following sections discuss possible shortcomings of our intervention design and measurement methods that might have influenced the study outcomes.
Aspects of the study design that may explain the lack of a noticeable interplay between exergame training and omega-3 FAs relate to the study population and to their intake dose and period. The selected older adults were quite fit and healthy elderly who had to be able to come to the study location by themselves. As all the participants showed a rather low level of omega-3 FAs at baseline, a huge potential existed to increase omega-3 FAs levels. The fish oil intake group showed a significant increase of the omega-3 FAs including DHA and EPA comparing pre vs post blood sample values. For fish oil supplementation, the intake duration of 16 weeks was long enough to reach a steady state condition [
88‐
90] and the intake amount was high enough to trigger a significant increase of the omega-3 FAs values in the blood. For the omega-3 index, all the participants within the fish oil group adapted from a undesirable level of less than 4% or an intermediate-risk zone of 4–8% to a cardio protective level of 8% or higher [
91] after 16 weeks. We concluded that all the participants responded well to the fish oil supplementation. From week 17 to 26, we noticed a steady state of the values while four participants showed a slight decrease of the omega-3 index to an intermediate risk zone (ranging from 5.15 to 7.81%). Even though the intake amount of 2.9 g omega-3 FAs per day seems to be an appropriate level, an individual adapted intake level might be even more promising because of genetic heterogeneity [
92]. The expectation was that these increased blood levels would enhance the effects of exergame training. However, we cannot directly link the blood value course to the integration and implementation of omega-3 FAs into the neuronal system as the efficacy is limited to use blood fatty acid levels as a surrogate biomarker for central nervous system levels [
93]. Therefore, the supplementation period of 16 weeks was long enough to reach a steady state level in the red blood cells of our study population, but we don’t know if the time period was also long enough to trigger an implementation effect into the brain cells and metabolism.
Another factor that might have influenced the interplay is the composition of the fish oil. The used fish oil contained DHA and EPA. However, the amount of EPA was higher than DHA. The brain contains high levels of DHA, but low levels of EPA [
94]. DHA is the component that is quantitatively the most important omega-3 FAs in the brain, having unique and indispensable functions in the neuronal membrane, and in turn has positive effects on the brain [
85,
93]. EPA has independent effects, particularly in regards to the respective anti-inflammatory mediators [
93]. In rats, EPA and DHA increased neurite outgrowth in the development stages, nonetheless only DHA triggered positive effects in the tissue of aged rats [
93]. Finally, studies indicated that the greatest benefits may be with DHA supplementation in non-cognitively impaired older people [
93]. Considering these potential shortcomings, we propose that future studies should choose the fish oil composition according to their expected effects. In our study, it can be hypothesized that a higher amount of DHA might have evoked stronger effects. It seems fair to state that a better understanding of the roles of DHA and EPA to support brain health and protection is needed [
93].
A further effect limiting factor could be related to the selected placebo, olive oil. We cannot exclude that the participants could profit from the olive oil as olive oil contains some effective components as well; for example, oleic acid [
95]. However, the focus of this study was on the effects of omega-3 FAs and the olive oil group showed no significant increase of the omega-3 FAs during the intervention. Therefore, we concluded that the effects evoked by omega-3 FAs were minimal in the olive oil intake group. For placebo, future studies could use a fish oil supplement with a different amount of DHA and EPA as the active examined supplement depending on the intended effects.
Some methodological aspects of the selected measurements might have also limited the possibilities to capture enhanced effects. To assess omega-3 FAs values, we were bound to blood sample analysis. As mentioned before, the efficacy to use blood fatty acid levels as a surrogate biomarker for central nervous system levels is limited [
93]. Moreover, it might be that our assessments, like measuring RLP, executive functions, and spatio-temporal gait parameters, were not sensitive enough to catch any effects at that stage. Neuroimaging methods might have provided more informative results than our more behavioral focused assessments. Furthermore, the used neuroimaging method TMS was probably too focused on the motor cortex only and not able to assess other brain areas. TMS measurement was limited to measure corticospinal excitability from the motor cortex to the right leg muscle (
M. tibialis anterior). We recommend, therefore, using brain imaging methods and protocols that are not restricted to a certain brain region. A recent systematic review of Tian and colleagues, that mapped relevant brain areas for gait variability, showed that several brain regions are important for gait performance [
96], and should, therefore, be considered in addition to the motor cortex. Brain imaging methods, e.g. magnetic resonance imaging (MRI) and positron emission tomography (PET), have the possibility to measure changes of grey and white matters as well as metabolic processes that might be better indicators for neuroplastic changes. A recent study, that also examined a combined approach including omega-3 FAs, stated that gray matter volume measurement might be more sensitive than behavioral outcomes to detect differences between a combined versus a single intervention [
86]. Furthermore, these imaging methods would allow the measurement of several brain areas and would be able reaching deeper-located brain regions, e.g. the hippocampus. Nevertheless, it remains open if our measurement methods were not sensitive enough to catch an effect or whether our study procedure was not designed appropriately enough to trigger an evident result.
Limitations
Some limitations of this study were already mentioned in the discussion section. In this section, we mention additional limitations that are not directly linked to the discussion section. Consideration of baseline fitness level as well as the change of the fitness level after the exergame training might have provided more information that could be used for the result discussion. Moreover, at the end of the intervention, we did not ask the participants about their assumption which nutritional supplementation they took during the intervention. Olive oil as placebo was chosen as it was the most similar to fish oil regarding to outer appearance and consistency. Nevertheless, this step could have substantiated the double-blind design. For the EEG measurements we used, some issues related to our measurement protocol might limit the measurement interpretation. The analysis showed a great range of latencies in the observed peaks. For example, the latencies (mean ± standard deviation) of the negative peak after response onset were for the fish oil and exergame group pre: Fp1 420.40 ms ± 243.13 ms, Fp2 385.60 ms ± 213.71 ms and post: Fp1 464.53 ms ± 219.58 ms, Fp2 501.26 ms ± 238.12 ms and for the olive oil and exergame group pre: Fp1 518.00 ms ± 254.31 ms, Fp2 514.00 ms ± 250.53 ms, post: Fp1 423.76 ms ± 231.23 ms, Fp2 431.13 ms ± 261.72 ms. It cannot be ruled out that differences in EEG cap positioning during the different measurement events are, at least in part, responsible for this observed variability. Nevertheless, the RLP shape was evident for the included participants, while in a few participants the RLP appeared at a later time point. Another reason for the shift in time might be due to a technical problem. Since the EEG activity was recorded using wireless signal transmission, it can be speculated that the signal transmission was slightly delayed in a few participants. Nevertheless, during the experiments other electrical devices were switched off to minimize interference. Furthermore, the randomized study design can be considered most optimal for controlling factors related to measurement issues.