The early twenty-first century confronts us with a dilemma concerning old age, as there is expected to be an unprecedented increase in the numbers of elderly people (Esiri and Chance, 2012). This is relevant because, e.g., the human brain experiences a variety of prominent structural changes during the course of aging (Deary et al., 2009; Scafato et al., 2010; Eggenberger et al., 2016), which often leads to a functional decline and disability (Borson, 2010). Structurally, the brain's size decreases with anterior to posterior progression, leaving the prefrontal regions most affected by this aging process. Increased age is also associated with a reduction of cognitive functioning, which is prevalent in almost every second elderly person (Scafato et al., 2010). Although the evidence appears not to be sufficient to conclude that modified brain structures reflect the neuroanatomical substrates for the age-related decline of cognitive performance (Fjell and Walhovd, 2010; Salthouse, 2011; Bennett and Madden, 2014; Eggenberger et al., 2016). This seeming discrepancy might be explained due to the ability of the aging brain to compensate for structural atrophy by amending recruitment patterns that support functioning and maintain cognitive functions (Deary et al., 2009; Grady, 2012; Eggenberger et al., 2016). Deterioration of the mental learning processes may affect a variety of cognitive domains, such as psycho-motor skills, (working-) memory, and executive functions (Skvarc et al., 2016). Executive functions are “higher-order aspects of cognition that are strongly dependent upon the integrity of the prefrontal cortex” (Lipnicki et al., 2009).

Working memory, being a vital process underlying human thought, is necessary for everyday decision-making and problem solving, making it a fundamental process in the daily lives of older adults (Nissim et al., 2016). It relies on frontal lobe structures and is known to decline with age (Nissim et al., 2016). Activities of daily living as organizing and planning daily routines and appointments require working memory and other components of the executive functions (Mograbi et al., 2014; Nissim et al., 2016). Thus, a decline in working memory capacity may lead to loss of independence and decreased quality of life (Klingberg, 2010).

Furthermore, cognitive decline may appear after surgery in 25% of the elderly population. Oxidative stress and adjacent neuro-inflammation of the brain are associated with cognitive decline, e.g., in post-operative delirium (Skvarc et al., 2016), however, the mechanisms of cognitive decline following surgery cannot be explained exactly for all elderly. The burden of post-operative cognitive deterioration is arduous, given the increased length of hospital sojourning, including prolonged stay on intensive care unit(s), and reduced functional performance; factors that not only have strong effects on the subjective way patients perceive the quality of their life, but also have an impact on the healthcare systems by increasing costs of care (Cropsey et al., 2015).

Cognitive training may reduce the risk of cognitive decline in elderly (Baumgart et al., 2015). A potential countermeasure to impede the cognitive decline is aerobic exercise-, as research showed that participation in aerobic exercise interventions improves functioning of the prefrontal cortex and performance of executive functioning tasks (Lipnicki et al., 2009). Improvement of the corticospinal excitability, a neurophysiologic mechanism related to muscle functioning, may also ameliorate the activity of the prefrontal cortex and executive functions of the brain (Kawakami et al., 2001). In a recent study with male adolescents, Benzing et al. reported that higher order cognition was immediately enhanced by cognitively engaging exergame-based physical activity (PA). A PA condition with a high level of cognitive engagement resulted in significantly better performance in cognitive flexibility compared to conditions with low levels of cognitive engagement (Benzing et al., 2016). Exergames are “technology-driven physical activities, such as, videogame play, that require participants to be physically active or exercise in order to play the game¹.” Both aerobic exercise and cognitive training in elderly participants are reported to improve brain health of the trainees, however, specificity of training seems to apply (Chapman et al., 2016). Chapman et al. (2016) showed that exercisers with greater gains in memory benefitted from higher central blood flow in the hippocampal regions, a brain area particularly vulnerable regarding aging related deterioration (e.g., dementia). These findings were consistent with a report of Shatil et al. (Shatil, 2013) where the authors proposed that cognitive training may be the main treatment for cognitive gains in elderly when compared to the benefit of isolated physical exercise.

More recent approaches showed that cognitive functioning may benefit from simultaneous cognitive-physical exercise programs, where traditional exercise components; e.g., aerobic and/or resistance training, are combined with video-game based cognitive training (Eggenberger et al., 2015). It can be hypothesized that a cognitive-physical training program delivered through exergames in elderly threatened with cognitive decline after medical treatment (e.g., surgery), shows effect on cognitive functioning (de Bruin et al., 2010). This is important since these elderly individuals have an impaired capacity to train adequately with aerobic types of training due to surgery related immobilization. Additionally, recent studies identified the potential for cognitive (motor-) training as a promising form of exercise for immobilized elderly patients in acute care settings (Pichierri et al., 2011; Schoene et al., 2014; van het Reve and de Bruin, 2014).

In accordance with the phased iterative approach suggested by Campbell et al. (2000) we designed and evaluated a bedside COGnitive and PHYsical training CONsole (COPHYCON) (Figure 1) for use in acute care patients (e.g., on intermediate and intensive care units). Before larger scale studies and application in clinical practice are to be considered, possible new treatments usually have to go through a series of phases to test whether they are safe and effective (Thabane et al., 2010; Bower et al., 2015). We, therefore, performed a trial including a single session of exergaming based on the model for complex interventions advocated by the British Medical Research Council (Craig et al., 2008). This was performed to test the usability and acute effects of the COPHYCON (Esiri and Chance, 2012) in terms of acceptance, adherence, and attrition (Lipnicki et al., 2009) and the acute effect of a COPHYCON exercise session on measures of brain functioning in a group of untrained elderly.

We hypothesized, that video game-play with the COPHYCON is usable and affects prefrontal cortical involvement.

Fifteen participants were included in this study (five women, 10 men) aged 63.7 (SD 5.4) years. Four men and one woman were randomized to NIRS measurements and four women and six men were randomized to EEG measurements, respectively. The demographic and baseline characteristics of the participants are reported in Table 1.

This study tested the usability and effects of the COPHYCON in terms of (1) acceptance, adherence and attrition and safety, and (2) the effect on measures of brain functioning in a group of untrained elderly adults. The findings revealed a high level of acceptance and high adherence rates. Evaluation of subjective usefulness suggests that the participants perceived COPHYCON based exergaming as useful to train while lying supine in a hospital bed. Moreover, participants expressed a positive attitude toward the program and a moderate behavioral intention to use it. The ease-of-use was rated high and no safety issues or device related adverse events were recorded during the test phase of the COPHYCON.

Playing the games had an effect on measures of prefrontal cortex functioning as assessed with fNIRS and EEG, confirming our initial hypothesis. This is in line with Eggenberger et al. (2016), who report long-term effects of video game play on prefrontal cortex behavior and function (Eggenberger et al., 2015). Compared to our study, this work differed in methodology, as this study performed fNIRS measures with another protocol. Nonetheless, Eggenberger (Benzing et al., 2016) reported both dance and balance training activities significantly reduced left and right PFC oxygenation during acceleration of walking. In contrast to the work of Eggenberger et al., our participants were positioned in supine position and the relative Hb02 level decreased more in the right prefrontal cortex while exergaming. These differences may be explained by functional changes of the brain in elderly people. The “Hemispheric Asymmetry Reduction in Older Adults-Model (HAROLD)” (Cabeza, 2002), provides a framework for the generalization of recall-, working-, and episodic-memory tasks in the elderly brain. Younger people reportedly have a stronger activation at the unilateral side of the brain. In contrast, in elderly the activation spreads to other parts of the brain, becoming bilaterally active (Cabeza, 2002). Berlingeri et al. (2013) however, reassessed the HAROLD model, concluding that this model only captured some of the brain patterns in aging. The results of this group were more compatible to the CRUNCH (Compensatory-related utilization of neural circuits hypothesis)-model, suggesting that former neurologic patterns can be considered a special manifestation of an age related compensatory process. The CRUNCH model posits that additional activations, regardless of the hemispheric side, are seen in elderly participants for comparatively easier versions of the same task. However, as the task demands increases, elderly would reach a sort of plateau corresponding to the highest possible level of activation. At this point, their performance would inevitably fall due to the impossibility of meeting the additional task demands (Berlingeri et al., 2013). This might be the case in the participants in our study, as participants reached a plateau in the Select games. Another concept of “brain plasticity” is the “cognitive reserve (CR)” theory. CR postulates that individual differences in the cognitive processes or neural networks underlying task performance allow some individuals to cope better than others with brain deduction and damage (Esiri and Chance, 2012), as there are individuals with a substantial load of pathology who are able to perform with small to moderate cognitive deficits within the normal span during lifetime (Dowling et al., 2010; Esiri and Chance, 2012). Such processes reflect “functional brain and cognitive plasticity” of the deteriorating human brain (Skvarc et al., 2016). We cannot exclude that our participants had no “functional deficits,” as the only performed functional test was the MoCA.

In contrast to the decreased oxygenized Hb02 levels in four of our participants, the level of Hb02 in one participant increased in both the left and right prefrontal cortex. It is conceivable that the other participants were challenged more during gameplay. Furthermore, the increase in oxygenated Hb02 might have been caused by the electrodes being positioned to far to the temporal area, or were due to intra-individual variations of the human brain (Allen et al., 2002).

“Behavioral Intention to Use” showed that participants were motivated to use the tested games on a regular basis. The moderate acceptance rate of 65% obtained in this study seems in line with a previous study (Laver et al., 2011) that revealed skepticism about the use of the COPHYCON and cognitive games on a hospital ward. A likely explanation may be found in the good physical and mental fitness of our participants and therefore, the reduced possibility of the games to challenge the participant's cognitive potential. The cognitive games used were developed for patients with cognitive impairments and deficits. Due to the training principle of initial values, improvement in the outcomes Alertness and Selective Attention will prevail most in patients with lower initial values. In other words, those with lowest level of cognitive fitness have greatest opportunity for improvements (Ammann et al., 2014). Future studies that want to assess the effects of training in a longitudinal study design should consider this effect.

Eggenberger et al. (2016) compared an 8-week interactive cognitive-motor video game dancing intervention with a conventional balance and stretch training in elderly (Eggenberger et al., 2016). Prefrontal cortex activity was assessed during walking on a treadmill. A decrease of the HB02% at preferred and fast walking speed, as well as an asymmetry of HB02% between both the left and right prefrontal cortex during treadmill walking was revealed, with larger effects in the video game dance group (Eggenberger et al., 2016). A major difference between the study of Eggenberger et al. and ours was the dose (intensity, frequency, and time) of the performed trainings. Participants in our study trained in a single intervention, while Eggenberger et al. subjected their participants to a training program of several weeks. For future studies, the understanding of the mediating effects on the brain, while performing a physical and/or cognitive training intervention is of interest to define the dosage of such interventions. Physical activity with high cognitive engagement is likely to be more efficient than physical activity of the same intensity with low cognitive engagement (Benzing et al., 2016). Thus, the effects of exergames with higher cognitive intensities may be also effective to improve acute cognitive function in elderly with non-communicable disease (Ammann et al., 2014; Stanmore et al., 2017). This might be of interest both for health care professionals and game developers, when designing and implementing principles of exercise training in applied cognitive games (Knols et al., 2016).

In our study, EEG recordings of theta power significantly decreased in the averaged ~0.25–0.75-time interval post-stimulus onset over the left PFC sensor across the cognitive game levels, from AS1 to SEL1 and SEL2. There were no significant changes in the contralateral PFC's sensor Fp2 as well as in any of the other FOI (Table 5).

As reported by Teo et al. (2016), these results may be of interest for patients at risk for cognitive decline or with neurological disease. Rossini et al. reported “an increase” (i.e., “slowing”) of slow EEG waves (e.g., Theta frequency band) in the elderly, as progressive dysfunction and loss of synaptic contacts and neuronal abnormal apoptosis characterize the physiological aging of the brain (Rossini et al., 2013). Deterioration of Theta EEG power over the PFC was also reported in patients receiving a delirium after thoracic surgery (van der Kooi et al., 2015).

This may be of particular concern for elderly who develop surgery-related cognitive decline. Electrophysiological predictors provide a non-invasive method to preselect individuals for whom training in complex task and skill improvement outside the training task environment are more likely to be successful the higher their EEG alpha power over frontal regions is (Mathewson et al., 2012). This indicates the importance of starting rehabilitation of cognitive functioning already in early stages; e.g., when patients are confined to be immobilized: The higher their frontal alpha power is at initiation of mobilization, the better learning and skill improvements in rehabilitation will be (Mathewson et al., 2012). Moreover, Theta EEG power recorded by two sensors over the prefrontal cortex appeared to be significantly lower than those of healthy age matched controls (van der Kooi et al., 2015), and, importantly, Theta EEG power over the PFC seemed to be influenced by exergaming in older adults (Schättin et al., 2016).

These findings might have implications for the early rehabilitation of elderly with cognitive changes (and eventually for elderly who scheduled for elective surgery), as excessive synaptic-neuronal loss and/or pathologic phenomena interfere with the functional and structural plastic remodeling of the interconnectedness of brain structures.

Training with the COPHYCON implies stimulation of different multisensory mechanisms. These multisensory mechanisms relate, in turn, to brain neuroplasticity. A healthy nervous system is able responding to intrinsic and extrinsic stimuli by reorganizing its structure, function, and connections; a response ability defined as neuroplasticity (Cramer et al., 2011). In the adult nervous system alterations in central and peripheral inputs together with behavioral experience may lead to reshaping of connectivity (Ganguly and Poo, 2013). COPHYCON training offers stimuli from an artificial environment that have to be centrally processed and hence transformed into a meaningful movement. Previous research has shown that the study of cortical responses in multimodal integration sites of the brain help in understanding the process of temporally accurate motor output based on the audiovisual cues and, furthermore, it is hypothesized that changes in these relationships can be monitored to gage performance changes in motor learning (Ono et al., 2014).

In this study, a usability design without a control group was used, as the results of the fNIRS and EEG measures are considered to be preliminary work. As postulated by Whitehead et al. (2014), “The distinction between feasibility and pilot studies is still a gray area, with various definitions having been suggested by clinical trial methodology researchers” (Whitehead et al., 2014). The end-points in this study are experimental for the time being and do not include clear criteria, to proceed to a main trial. Therefore, control participants were not included, but their inclusion should be considered in future projects. In such future projects, an important research focus relates to COPHYCON training, multisensory mechanisms, and brain plasticity in individuals with surgery related immobilization. A first study has shown this to be feasible (Turon et al., 2017). As interventions aimed at maintenance of “brain activity” in the elderly has consequences on everyday life (Rossini et al., 2013), measures of independent living and mobility should also be included (Posner, 2008). To evaluate the effect of exergaming on functional outcomes, e.g., attention, should also be assessed (Spence and Driver, 1997).

Several possible limitations of the study should be noted. First, this usability study included a user-centered interaction design with a single exergame-session. A longitudinal trend (either improvement or deterioration) in the acceptance of the COPHYCON, including the applied games or changes in the fNIRS/EEG outcomes (either improvement or deterioration) could not be determined. Second, fNIRS and EEG measures were performed with two separate devices. The results of the report of Leamy and Ward (2010) indicate that performing combined fNIRS/EEG measures should be preferred over using fNIRS or EEG devices separately, as combined FNIRS and EEG devices provide higher spatial resolution and higher temporal resolution (EEG) outcomes. Third, artifacts in fNIRS measurements are common. A difference between the measurements of the left and right sensor was reported and an asymmetry from 2 to 4% of the relative oxygen values may also occur (Casati et al., 2006). To date, there is limited guidance for clinicians to compare the accuracy of fNIRS measures, due to the possibility of the variability in methodologies, bias in patient selection, and among fNIRS devices itself. The device used in this study was able to measure the absorbed light by the electrodes 10 times per second. An average for rS02% of 50 data points was calculated, meaning that a value was shown on the fNIRS-screen every 5 s (Thavasothy et al., 2002). This procedure was not problematic in this study design, as we observed rS02%-values over 80 min; however, if the aim had been to detect fast changes in rS02 on events during cognitive gameplay, an evaluation of the efficacy of the stimuli in the games would not have been possible. Therefore, the measures and results of the INVOS 5100c device in this study cannot be compared directly with other fNIRS measures (Hyttel-Sorensen et al., 2014; Seiyama et al., 2016). Fourth, rS02% of the prefrontal cortex. Tachibana et al. (2011) performed Hb02 measures of the superior parietal lobe and superior temporal gyrus in participants performing a dance game, in which the oxygen values increased, as the game advanced in difficulty. This is relevant, as differences in the temporal relationships of the parietal lobes and the superior temporal gyrus may reflect the functional roles of the mean oxygenized hemoglobin concentration levels while performing video training. Further studies including fNIRS measures with several videogames as indicator for motor learning should therefore also measure the temporal areas, if possible. Fifth, the only neuropsychological test which was performed in advance of the exergame training was the MoCA. However, before the participants began the COGNIPLUS training, the software of the CogniPlus system demanded an initial stage-test, in which the performance level of the participant could be determined. After this stage test, the software ran a higher or lower level depending on the errors and the adequate responses, as well as to the response time of the required task during the game, e.g., a good response to a car (suddenly) crossing or not crossing the street. If the performance of a participant appeared to be better than the game level, then the level increased with one stage. Contrary, if the performance decreased, the software automatically decreased the stage to a lower level. Last, our participants were ambulatory community dwelling elderly in good mental and physical health. Their preferences for computer games and device use in rehabilitation may be different to those, whose health state depends on rehabilitation, e.g., long stay patients at risk for cognitive and functional deterioration. This might be challenging as alertness, attention and perception of these patients may vary and the time span of concentration may be short.

This study featured a rather small sample of 15 healthy elderly, which resulted in limited statistical power for fNIRS and EEG analysis and, thus, the validity of using elderly persons to test the usability may be questioned. However, the primary focus of this study was on usability, where a sample of five participants is sufficient for detecting 80% of the usability problems (Lewis, 1994). Further studies should determine the effect of studies with adequate samples.

Usability studies provide invaluable results for information technology developers and rehabilitation specialists who wish to use innovative systems. Findings of such studies are relevant to the development of new emerging virtual reality-based rehabilitation methods that try to enhance prefrontal cortical involvement in elderly. This study demonstrates that the game-based intervention was perceived as usable and had a positive effect on prefrontal cortical involvement. Further testing of the device in a clinical surrounding is warranted.

The dataset analyzed in the current study is available from the corresponding author(s) on reasonable request.

RK: conceived the methodology and carried out quality assessment, data analysis, and manuscript writing; JS: participated in methodology conception, quality assessment, and manuscript writing; DDB: carried out data collection, quality assessment, data analysis, and manuscript writing; FG: was technical advisor for the EEG technology, carried out EEG data analysis, and participated in manuscript writing; MW: was technical advisor for the fNIRS technology and participated in manuscript writing; BK: carried out fNIRS data analysis and manuscript writing; DB: was responsible for the quality assessment and manuscript writing; EdB: supervised progress, helped with methodology conception, manuscript writing, and critical revision for scientific content; All authors read and approved the final manuscript.