Lifelong physical activity and executive functions in older age assessed by memory based task switching

Author’s Accepted Manuscript Lifelong physical activity and executive functions in older age assessed by memory based task switching Patrick D. Gajewski, Michael Falkenstein www.elsevier.com/locate/neuropsychologia

PII: DOI: Reference:

S0028-3932(15)30013-0 http://dx.doi.org/10.1016/j.neuropsychologia.2015.04.031 NSY5577

To appear in: Neuropsychologia Received date: 3 February 2015 Revised date: 28 April 2015 Accepted date: 29 April 2015 Cite this article as: Patrick D. Gajewski and Michael Falkenstein, Lifelong physical activity and executive functions in older age assessed by memory based task switching, Neuropsychologia, http://dx.doi.org/10.1016/j.neuropsychologia.2015.04.031 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Lifelong physical activity and executive functions in older age assessed by memory based task switching Patrick D. Gajewski* & Michael Falkenstein Leibniz Research Centre for Working Environment and Human Factors (IfADo), Dortmund, Germany

* Corresponding author: Patrick D. Gajewski Leibniz Research Centre for Working Environment and Human Factors at the Technical University of Dortmund, Ardeystr. 67 D-44139 Dortmund, Germany Telephone: ++49 231 1084 383 Fax:

++49 231 1084 308

Email: [email protected] URL: http://www.ifado.de Running title: Executive functions and long-term physical activity in aging. Abstract Aging is accompanied by compromised executive control. Training studies point to beneficial effects of physical activity on executive functions. Here, we investigate the relationship between lifelong habitual physical activity (about 50 years) and switch ability in healthy seniors. Participants switched among three tasks in a memorized task sequence. Mixing costs for speed were lower in habitually active than low active participants whereas switch costs were not affected. Active participants revealed also lower mixing and switch costs for accuracy. These parameters were negatively correlated with the self-reported level of physical activity. The frontal CNV was smaller in the active than low active group. In contrast, in the target-locked ERPs active individuals showed an earlier P2, a larger frontocentral N2 and the typical pattern of smaller P3b in switch than non-switch trials relative to low-active individuals. These data suggest that lifelong physical activity is associated with faster recall of stimulusresponse sets (P2), enhanced response selection during interference processing (N2) and working memory updating (P3b) leading to lower mixing and switch costs.

Keywords: ERP, Aging, task switching, physical activity, P2, N2, P3b 1. Introduction Goal directed behaviour is crucial for humans to act purposefully in changing environments. Executive functions enabling goal directed behavior are managed by dynamic neural systems interacting continuously with environmental requirements (Diamond, 2013; Miyake et al., 2000). Aging is associated with an impairment of some executive functions due to an age-related decline of those systems (Salthouse et al., 2003; West, 1996). However, a growing body of evidence suggests beneficial effects of physical activity and particularly aerobic exercise on executive functions in elderly (Colcombe & Kramer, 2003; Etnier & Chang, 2009; Hillman et al., 2008; Kramer & Erickson, 2007a,b; Gomez-Pinilla & Hillman, 2013; Guiney & Machado, 2013; Voelcker-Rehage & Niemann, 2013, for meta-analyses and reviews). Most of the reports analyzed effects of physical activity on executive functions in controlled intervention studies with interventions lasting up to 24 months (e.g. Colcombe et al., 2003; Hayes et al., 2013; Smith et al., 2010). On the other hand, there are some studies examining benefits of habitual but timely limited physical activity on executive functions in elderly using cross-sectional designs (Guiney & Machado, 2013 for review; Berchicci et al., 2013; Chang et al., 2010; Taddei et al., 2012; van Boxtel et al., 1997; Wendell et al., 2014). In sum, current research indicates that short- and middleterm physical activity counteracts age-related decline of executive functions (Colcombe et al., 2003, 2004, 2006; Draganski & May, 2008; Erickson et al., 2010, 2011; Flöel et al., 2010; Gomez-Pinilla & Hillman, 2013; Guiney & Machado, 2013; Hayes et al., 2013; Hillman et al., 2008; Voss et al., 2010; 2011; 2013; Weinstein et al., 2012). Therefore, long-term or even lifelong physical activity should attenuate agerelated executive deficits to a much larger extent. 2

However, there is to our knowledge no study which examined explicitly the relationship between lifelong (across several decades) physical activity and executive functions and their neural correlates in older individuals.

1.1 Task switching in older age An important executive function, supporting goal-directed behavior is the ability to switch among different tasks while maintaining a goal. A tool to investigate and to separate different functional components during task switching is the memory based task switching paradigm (Rogers & Monsell, 1995). Task switching is not a unitary concept but consists of a number of cognitive components (Kiesel et al., 2010) which are differentially affected by age (Cepeda et al., 2001). Recent studies showed deterioration of specific sub-mechanisms involved in task switching in aging (Butler & Weywandt, 2013). For instance, age-related changes have been shown during task preparation and during maintaining multiple task sets in working memory (Cepeda, et al., 2001; Gajewski et al., 2010b; Kramer et al., 1999; Kray & Lindenberger, 2000; Mayr, 2001; Mayr & Liebscher, 2001; Meiran et al., 2001; Kray et al., 2004; Kray, 2006; West & Travers, 2008). Most researchers have been using an experimental task design consisting of blocks with single tasks only and blocks with task switching (mixed blocks). This allows the computation of two types of switching costs: mixing (or global) and switching (or local) costs. Mixing costs are defined as performance differences between single task blocks and task repetition trials in mixed blocks. They are assumed to reflect updating and working memory processes (Kray et al., 2010) or the resolution of task ambiguity in the mixed blocks (Los, 1996; Mayr, 2001). Switching costs are defined as performance differences between switch and non-switch trials in mixed blocks. They are usually associated with interference between conflicting task sets (Allport et al., 1994; Kiesel 3

et al., 2010). Mixing costs are usually larger in old than in young adults while switching costs are not (Cepeda et al., 2001; Karayanidis et al., 2011; Kramer et al., 1999; Kray & Lindenberger, 2000; Kray et al., 2005; Kray, 2006; Wasylyshyn et al., 2010 for review). The age-related increase in mixing costs is particularly evident in memorybased switching when the trial sequence has to be continuously maintained and the currently relevant task-set retrieved from memory (Kray & Lindenberger, 2000; Reimers & Maylor, 2005; Schapkin et al., 2014).

1.2 Aging effects on ERP components during task switching Event related potentials (ERPs) offer deeper insights in the functional structure underlying task switching. Since ERPs have an excellent time resolution, they allow analysing each function involved in task switching separately (Jost et al., 2008). In particular, task-set updating as reflected in the P3 component after cues is attenuated or delayed in older age (Friedman et al., 2007; Karayanidis et al., 2011; Kray et al., 2005; West, 2004; West & Travers, 2008). The findings regarding the CNV indicating anticipatory attention and task preparation are less consistent (Brunia & van Boxtel, 2001; Falkenstein et al., 2003; Walter et al., 1964). In some studies the CNV was reduced for older vs. young subjects (Sterr & Dean, 2008; West & Moore, 2005), but it has been also found to be larger in older than younger individuals in effortful conditions (i.e. under time pressure), suggesting stronger attention/preparation in elderly to maintain a sufficient level of performance (Berchicci et al., 2012; Goffaux et al., 2008; Karayanidis et al., 2011; Kray et al., 2005; Wild-Wall et al., 2007). After the target onset the cognitive processes related to stimulus encoding, retrieval of S-R mapping and response selection and execution have to be completed. Each of these processes has been associated with a particular ERP component. In context of 4

task switching the target-locked frontal P2, frontocentral N2 and parietal P3b have been studied. The P2 has been related to the retrieval of stimulus-response associations (Adrover-Roig & Barceló, 2008; Gajewski et al., 2008; Kieffaber & Hetrick, 2005). Schapkin et al. (2014) found a larger P2 in older than younger participants in mixed blocks, whereas no age difference was found in single task blocks. According to the proposal of Finke et al. (2011) an increase of the P2 amplitude may reflect an additional effort of target processing in difficult conditions like switch trials. This was corroborated by a positive correlation between the P2 amplitude and RTs, suggesting the larger the P2 amplitude the slower the response. Moreover, Schapkin et al. (2014) found that the P2 latency was positively correlated both with reaction times and mixing costs in RTs in older participants. These findings support the idea of S-R retrieval of the P2 (Kieffaber & Hetrick, 2005) as the timing of the P2 peak is related to the timing of the correct response. Following the P2 a large negative peak is visible, the frontocentral N2. The N2 was primarily associated with detection of mismatch and cognitive response monitoring (Folstein & van Petten, 2008; Hämmerer et al., 2014 for reviews), conflict processing (van Veen & Carter, 2002; Yeung & Cohen, 2006), and more generally to decision making (Ritter et al., 1979, 1982) and response selection i.e. the implementation of SR mappings (Gajewski et al., 2008, 2010a; 2011). Response selection is intensified and delayed when response conflict has to be resolved, which is often reflected in enhanced and delayed N2 in switch vs. non-switch trials (Gajewski et al. 2010a). The N2 is generally reduced and delayed with increasing age (Friedman, 2008; Hämmerer et al., 2014 for reviews), which suggests impaired response selection and/or compromised conflict resolution. The subsequent target-locked P3b has been repeatedly described in context of task switching. The common pattern is a lower P3b-amplitude in switch vs. non-switch trials 5

(Barceló et al., 2000; Gajewski & Falkenstein, 2011; Lorist et al., 2000; Rushworth et al., 2002). Moreover, the P3b is consistently larger in single task blocks than in mixed task blocks (Gajewski & Falkenstein, 2012; Jost et al., 2008; Karayanidis et al., 2011). Generally, the P3b amplitude decreases and its latency increases with age (e.g., Friedman, 2008; Polich, 1996).

1.3 Association between aerobic exercise and ERPs It has been shown that not only performance but also some ERP components are modulated by physical activity (Gomez-Pinilla & Hillman, 2013; Guiney & Machado, 2013). For example Hillman et al. (2002) as well as Kamijo et al. (2010) showed that physically fit participants have smaller CNV amplitudes than non-fit individuals, suggesting generally less preparatory effort and higher preparation efficiency in physically active persons. To our knowledge, only few studies investigated the N2 in older active versus inactive individuals. Taddei and colleagues (2012) who investigated young and older fencers and non-fencers in a go/no-go task reported faster reaction times, larger and earlier N2 both in young and in older fencers vs. non-fencers, supporting previous findings investigating young fencers only (Di Russo et al., 2006). In contrast, no association between exercise and the N2 was found in young adults subdivided in an aerobically active vs. low active group on the basis of maximal oxygen consumption (Themanson et al., 2006a). These discrepant results may be due to different types of physical activity and/or different tasks. Studies investigated associations between exercise and the P3b showed shorter latency and larger amplitude in physically active than low active adults in stimulus discrimination tasks (Polich & Lardon, 1997; Pontifex et al., 2009).

6

1.4 Association between aerobic exercise and ERPs during task switching There are only few reports showing the relationship between physical activity and task switching ability using ERPs. Kamijo and Takeda (2010) using an alternative runs paradigm (Rogers & Monsell, 1995) found smaller switching costs in young physically active vs. inactive participants but no differences in the ERPs. Scisco, Leynes and Kang (2008) did not find any differences in task switch performance and ERPs between young physically fit and non-fit participants. The authors concluded that the relationship between greater cardiovascular fitness and improved cognitive function may emerge only later in adulthood. Indeed, Hillman et al. (2006) found lower mixing and switching costs and shorter latencies and larger amplitudes of the P3b in physically active than inactive older individuals. Themanson et al. (2006b) reported lower mixing costs but not switching costs in active vs. inactive older adults. More recently, Dai and colleagues (2013) compared three groups of older adults with openskills (e.g. tennis), close-skills (e.g. jogging) and irregular physical activity. The total duration of the physical activity between groups varied between 11 and 13 years. The open skill groups revealed the lowest mixing costs, following by the closed-skill and the no activity groups. No group differences in local switching costs were obtained. The ERPs showed larger P3b amplitudes in both active groups supporting the findings of Hillman et al. (2006) but no further ERP components were analyzed.

1.5 The present study The aim of the present study is to elucidate behavioral and several electrophysiological indices of executive functions associated with lifelong physical activity. To this end, we compared performance and event-related potentials (ERPs) in a task switching paradigm between long-term (about 50 years) physically active and low-active healthy, elderly men. 7

Most of the reported ERP studies used a task cuing paradigm (e.g. Jost et al., 2008) allowing analysis of task set updating and maintenance evoked by the cue during task preparation. Some studies used the alternative-runs paradigm mentioned above, which requires switching among two tasks in a fixed order (AABBAA etc.) but this version is only partly memory based as the relevant task was additionally indicated by the spatial position of the stimulus (Rogers & Monsell, 1995). The present study does not use explicit cues. Instead, a fully memory-based task set retrieval was necessary to perform the task. We chose this version for creating a task which is particularly difficult for older subjects due to the increased memory load. Therefore, the current study offers a unique combination of lifelong physical activity and task switching ability in a fully memory based task switch paradigm. Furthermore, we were interested in electrophysiological underpinnings of behavioral differences between physically high and low active adults. There is evidence that superior performance in executive control tasks is accompanied by larger ERPs due to the recruitment of frontal regions (Prakash et al., 2011). This enhanced activity related to superior performance is usually evident in larger ERP amplitudes and/or shorter latencies in, for example, younger vs. older adults (Friedman, 2007; Hämmerer et al., 2014), in physically active vs. inactive older adults (Dai et al., 2013; Dustman et al., 1990; Getzmann et al., 2013; Hillman et al., 2006) or after vs. before cognitive training regimes (Gajewski & Falkenstein, 2012; Wild-Wall et al., 2012). Thus, if physical activity confers a benefit in executive control processes, we expect active elderly to exhibit lower mixing and/or switching costs on a behavioural level, as compared to low active individuals. Moreover, we expect a lower intra-individual variability of speed quantified as individual standard deviations (ISDs) in the former group as it has been shown that this parameter reflects individual differences in neuronal integrity (Hultsch et al., 2000; Hultsch & MacDonald, 2004), which can be 8

enhanced by aerobic fitness at least in children (Moore et al., 2013; Wu et al., 2011). To date, no such finding has been reported for older adults. On the neurophysiological level we predict mainly effects on ERPs associated with preparatory activity, indicating by the CNV. The task set updating process reflected in the P3b is less synchronized in purely memory than cue-based switching. Thus, we expect that the P3b amplitude will be blurred in the preparation interval and should not differ between the groups of participants. In the target-locked ERPs we expect effects on S-R retrieval as reflected by the P2, response selection as reflected in the N2 and the P3b related to working memory and allocation of cognitive resources to the relevant task. In particular, we predict a lower CNV, a larger and earlier P2, and an enhanced N2 and P3b in physically high vs. low active old subjects.

2. Material and Methods 2.1. Participants Forty healthy male volunteers aged 65 to 87 (M = 73.2, SD = 4.5) participated in the study. Half of them represented a physically active group, while the other half a low active group (see next section). The active subjects were recruited from a local sports club; the low active subjects were selected from a large group of 152 low active healthy elderly with the aim to match the active seniors as closely as possible for demographic and lifestyle characteristics, as described below. As the sample was relatively small, we conducted a power analysis using Cohen’s d (Cohen, 1988) as assessed by ‘‘G*Power’’ (Faul et al., 2007; www.gpower.hhu.de). The power of a test estimates the probability of correctly rejecting the null hypothesis when it is false. We computed test power using a model for difference between two groups basing on the mean group differences between closed-skill (n = 16) and 9

irregular exercise (n = 16) groups for mixing costs and switch costs in reaction times using the results reported by Dai et al. (2013). The computed effect size for mixing costs was d = 0.53 which corresponds to a medium effect size, for the switch costs d = 0.01 which corresponds to a very small effect size. The estimated power for mixing cost was 0.42 and for switch cost 0.05. Therefore, we assume for our sample of n = 20 in each group a power of at least 42% for mixing costs and a low power of 5% for switch costs. The groups were matched regarding gender, age, education, no-smoking status, no history of cardiovascular, psychiatric, neurological, motor or oncologic diseases and psychopharmacologic or hormonal therapy, current or history of alcohol or drug addiction (see Table 1 for some demographic characteristics of the sample). Two participants of each group were left handed. All participants had normal or corrected-to-normal vision. The participants received payment for their participation. All participants gave written informed consent before start of the experiment.

2.2. Measurement of the physical activity All participants underwent a physical assessment consisting of a questionnaire designed by the authors for the purpose of this study to assess long-term physical activity, a standardized questionnaire measuring physical activity in the last two years, and a bicycle ergometry test measuring current physical fitness. In the questionnaire designed for the purpose of this study, we included questions on the age that participants started with regular physical activity, total duration, intensity, and type of sport. Participants in the low active group reported no current or history of regular physical exercise. In contrast, participants of the active group reported habitual long-term nonprofessional activity such as team games (volleyball, football), tennis, swimming, 10

fencing, rowing, jogging, gymnastic, strength training, or cycling beginning in adolescence. Thus, according to the self-report the active participants had been physically active on average for 50 years (SD: 13 years, range: 30-74 years). Six participants reported some discontinuities or fluctuations in the intensity of physical activity across their life, primarily due to time constraints in midlife. In addition, the amount of physical activity was estimated using an adapted version of the Lüdenscheid Activity Questionnaire (Lüdenscheider Aktivitätsfragebogen, Höltke and Jakob, 2002). The questionnaire is aimed at assessing physical activity and weekly energy metabolism due to sport activities as indices for risk factors for cardiovascular diseases. The questionnaire consists of 13 questions, such as, “How long do you walk per week?”, “How often do you climb stairs?”, “How long do you cycle per week?”, “How frequently did you exercise which type of sport in the last 2 years?” or “How much time do you spend on physical exercise per week?”. In general, the questionnaire provides an assessment of current physical activity and activity in the last 2 years. The questionnaire indicated a significantly higher level of physical activity in the physically active than in the low active group (Tab. 1). The members of the physically active group reported spending an average of 290 minutes on physical activity in 4.5 sessions per week (~1 hour per session). Finally, in the physical fitness session, participants’ current physical performance was measured with a bicycle ergometer using a physical work capacity (PWC-130) cycle test. The PWC-130 is a version of the PWC-150 test adapted for elderly individuals (Cambell et al., 2001). The aim of this test is to predict the absolute power output at a projected heart rate of 130 beats per minute. A relative power output was calculated by the power-to-weight ratio. As expected, there were highly significant differences in physical fitness between the active and low active group, both in the maximum power and in power-to-weight ratio (Table. 1). 11

The results of the activity questionnaire were in line with the ergometry assessment, with significant correlations between self-reported activity level and absolute power (r = 0.65; p < 0.001) and power-to-weight ratio (r = 0.69; p < 0.001). ---------------------------------Insert Table 1 about here ----------------------------------

2.3. Neuropsychological testing The participants underwent extensive neuropsychological assessment to document the neuropsychological and psychiatric status. The neuropsychological tests and the questionnaires were administered in an extra session, one day before the ERPs test was conducted. The battery included standardized tests and questionnaires that assessed general cognitive status (Mini Mental State Examination (MMSE); Folstein et al., 1975) and “Big Five” personality traits using the NEO-FFI (Costa & McCrae, 1992). The neuropsychological tests measured attentional endurance (d2; Brickenkamp, 1994), speed of processing and vigilance (Digit-Symbol-Test), short-term and working memory functions (Digit-Span-Test), both being subtests of the Wechsler Adult Intelligence Scale (Wechsler, 1956). Interference was assessed by the classic Stroop colour-word test (Stroop, 1935). Verbal memory was assessed by the Verbal Learning and Memory Test (VLMT; Helmstaedter & Durwen, 1990). Divergent thinking was measured by the German version of the Word-Fluency-Test (WFT; Aschenbrenner et al., 2001) and the crystallized intelligence was examined by the multiple choice wordtest (MWT-B; Lehrl, 1995). Mental rotation was assessed by the mirrored figures, a subtest from the Wilde Test of Intelligence (Jäger & Althoff, 1994). Finally, the Trail Making Test (TMT; Reitan, 1992) was administered to measure the psychomotor speed

12

a task switching. The self-reported failures in perception, memory, and motor function were analyzed using the Cognitive Failures Questionnaire (CFQ; Broadbent et al., 1982).

2.4. Stimuli and Tasks Stimuli in the task switching paradigm consisted of the digits 1 – 9, excluding the number 5. The digits were presented in white on the black computer screen 3 mm above the white fixation point (10 mm diameter). Each digit was presented in small (7 x 10 mm) and in large (12 x 18 mm) size on the computer screen. A cue (16 x 32 mm) indicating the relevant task was presented 3 mm below the fixation point. The cue “NUM” (German “Numerisch”, numeric) indicated a numerical task (greater or less than 5), “GER” (German “Geradzahligkeit”, parity) the parity task (odd vs. even), “SCH” (German “Schrift”, font) the font-size task (small vs. large). Explicit cues were presented in the single tasks only. In the memory-based mixed block (see below) three letters X were presented instead of the informative cue. Responses consisted of pressing one of two buttons which were mounted in a response box. The buttons should be pressed with the index fingers. The stimulus-response mapping of the three tasks was overlapping, that is responses according to ‘smaller than five’, ‘even’ and ‘small size’ were assigned to the left key and ‘larger than five’, ‘odd’ and ‘large size’ to the right key. The mapping was the same for all participants.

2.5. Procedure A schematic example of a trial in a single and in a mixed block is shown in Figure 1. ------------------------------Insert Fig. 1 about here. -------------------------------13

A trial started with a presentation of the fixation point. In the single task blocks a cue stimulus was presented for 1300 ms which remained visible when the target was presented. A response had to be given within 2500 ms after target-onset. 500 ms after the response a feedback stimulus was displayed for 500 ms. In case of a correct response a plus sign, after a wrong response a minus sign was displayed. 300 ms after the feedback stimulus disappeared the next (dummy) cue was shown. The response-cue interval (RCI) lasted in total 1300 ms and included the responsefeedback delay, feedback and the feedback-cue delay. At the beginning of the session participants preformed three single task blocks with a fixed task NUM, GER and SCH consisting of 34 trials each. The single task blocks were used to become familiar with the different task rules before the mixed block was run and to assess baseline performance in absence of a concurrent task. In the mixed block, the three tasks were presented in mixed order and the participants were instructed to switch the rule after every three trials in the following order “NUM– NUM-NUM-GER-GER-GER-SCH-SCH-SCH” etc. while a dummy cue “XXX” instead of an explicit cue was presented, i.e. participants had to keep the trial sequence in mind. When three consecutive errors were made or no response within the 2500 ms interval was given, explicit cues were presented for the next three trials, helping the participants to find the track. Additionally, the participants were instructed not to press the response button when they lost track of the sequence. The mixed block consisted of 126 trails. The frequency of task switch was 33.3%. The mixed block included an equal number of tasks rules (NUM, GER, SCH) targets (i.e. digits) as well as responses (i.e. left, right). The participants were given a written instruction that explained the task. A practice block consisting of 16 trials was used. The instruction encouraged quick and accurate responses.

14

2.6. Electrophysiological recordings EEG was recorded from 32 scalp electrodes according to the extended 10-20 system (Jasper, 1958) and mounted on an elastic cap. The montage included 8 midline sites and 12 sites on each hemisphere and two mastoid electrodes (M1 and M2). The EEG was re-referenced offline to linked mastoids. The horizontal and vertical EOG was recorded bipolarly from electrodes at both eyes. The horizontal and vertical EOG was measured by electrodes placed at the outer canthi (LO1, LO2) and above and below both eyes (SO1, SO2, IO1, IO2). Eye movement artifacts were corrected using the correction algorithm of Gratton, Coles and Donchin (1983). Electrode impedance was kept below 10 k. The amplifier bandpass was 0.01 - 140 Hz. EEG and EOG were sampled continuously with a rate of 2048 Hz. Offline, the EEG was downscaled to a sampling rate of 1000 Hz and cut in stimulus-locked epochs by using the software Vision Analyzer 1.05 (Brain Products, Munich). Epochs in which the amplitude exceeded +/- 150 µV were rejected. The ERPs were filtered digitally offline with a 17 Hz low and 0.05 Hz high pass.

2.7. Data Analysis The first trial of each test block, trials with responses faster than 100 ms or slower than 2500 ms, as well as error trials, were excluded from the RT analysis. RTs, standard deviations of RTs as an index of intraindividual variability of speed (ISDs) and error rates were subjected to two ANOVA designs: in the first design mixing effects were analysed by comparing mean performance of the single task blocks and the performance in non-switch trials of the mixed block. In the second design local effects were analysed by comparing non-switch and switch trials of the mixed block. Thus, the design 1 included a within-subject factor Block (single vs. mixed) and the betweensubject factor Group (active vs. low active). Design 2 included the within-subject factor 15

Task-Set Transition (non-switch vs. switch) and the between-subject factor Group (active vs. low active). In case of a significant interaction, a separate ANOVA was conducted. Group differences were assessed using one-way ANOVA.

The electrophysiological analysis was conducted for the (dummy) cue-locked and target-locked ERPs. Peak amplitudes and latencies of transient components were measured at their local maximal or minimal amplitudes in predefined time windows. Long lasting components (like P3b) or slow potentials (like CNV) were quantified as mean amplitude in a particular time window. The electrodes were a priori selected to reduce data complexity according to the potential distribution in the topographical maps. According to the theory-driven analysis of behavioral data, the ERPs were also analysed in terms of mixing effects and local effects. The timing of the single and mixed blocks was the same, enabling a direct comparison. The (dummy) cue-locked data was measured in the time window 0 – 1300 ms relative to the baseline -100 – 0 ms before cue onset. In this preparation phase following ERP components were measured: P2 at Fz as peak amplitude in the time window 150 – 300 ms, the N2 as peak amplitude in the time window 200 – 500 ms at Cz, P3b as mean amplitude 300 – 800 ms at Pz. The frontal CNV was measured as mean amplitude in the time range 1000 – 1300 ms at Fz. Target-locked ERPs were measured in the time window 0–1000 ms after target onset relative to the baseline -100 – 0 ms before target onset. Similar to cue-locked ERPs, the P2 was analysed at Fz as peak amplitude in the time window 150 – 300 ms and the N2 was measured as the most negative peak at Cz in the time range 200 – 500 ms after target onset. The P3b was measured as the mean amplitude at Pz in the time range 400 – 700 ms after target onset. For the P3b we used the mean amplitude

16

measure as the peaks cannot be always unequivocally detected. Only correct trials were analyzed. Accordingly to the statistical analysis of mixing effects in the behavioral data the ERP analysis included the factors: Block (single, mixed) and Group (active vs. low active). Local effects were analyzed by including following factors Task-Set Transition (nonswitch vs. switch) and Group. To test specific effects or interactions, additional oneway ANOVAs were employed. The relationship between behavioral and physical activity related parameters was assessed by correlation analysis using the Pearson correlation coefficient (two-tailed). The reported correlations were Bonferroni corrected. The numbers that follow the ± sign reflect standard errors (SEM). Effect sizes were computed to provide a more accurate interpretation of the practical significance of the findings, using the partial eta-squared coefficient (2). The statistical analysis was conducted by IBM SPSS 21.0.

3. Results 3.1. Neuropsychological data The sample was examined using neuropsychological tests and personality parameters. The results are given in Table 2. No relationship between physical activity and performance was found in almost all cognitive tests. The exception was the interference index of the Stroop test computed as the difference between the times to complete the interference and the colour naming list (list 3 – list 2). This interference index was significantly higher for nonactive than active participants (F(1,38) = 5.5, p < .05). Regarding the personality factors assessed by the NEO-FFI, only the dimension agreeableness was found to be higher in the active than in the non-active seniors (F(1,38) = 9.7, p < .005). 17

---------------------------------Insert Table 2 about here ----------------------------------

3.2. Behavioural data For the analysis of reaction times, error trials (2.1 % and 8.7 %) and outliers with RTs shorter than 100 ms and longer than 2500 ms (0.4 % and 5.3 %) for single and memory-based blocks, respectively, were discarded. Mean reaction times (RTs) individual standard deviations (ISDs) and error rates (ERRs) are presented in Figure 2. ------------------------------Insert Figure 2 about here -------------------------------3.2.1. Reaction times 3.2.1.1. Mixing effects The ANOVA for RT assessing mixing effects yielded a main effect of Block due to much longer RTs in the mixed than single task block (976 ± 30.0 vs. 644 ± 13.9 ms; F(1,38) = 180.5, p < .0001, η² = .826) and a strong trend for the effect of Group (F(1,38) = 3.9, p = .055, η² = .093), indicating slower responses in low active (850 ± 28.1 ms) compared to active seniors (771 ± 28.1 ms). Importantly, there was an interaction Block x Group (F(1,38) = 6.1, p < .05, η² = .138). This interaction was due to slower RTs in the mixed task block in the low than highly active group (1046 ± 38.1 vs. 906 ± 46.5 ms; F(1,38) = 5.4, p < .05), whereas no group difference was found for single task blocks (653 ± 24.1 vs. 635 ± 13.9 ms; F < 1). This pattern suggests larger

18

mixing costs in low active compared to physically active individuals (394 ± 31.2 vs. 271 ± 38.4 ms, p < .05). Expectedly, the individual standard deviations ISDs were larger in the mixed than single task block (362 ± 14.8 vs. 148 ± 6.2 ms; F(1,38) = 201.2, p < .0001, η² = .841). Moreover, there was a significant effect of Group (F(1,38) = 6.0, p < .05, η² = .137) due to larger speed variability in low active compared to active seniors (276 ± 12.0 vs. 234 ± 12.0 ms) and a trend for the interaction Block x Group (F(1,38) = 3.4, p = .07, η² = .082) due to the tendency for larger mixing costs in ISDs in low active compared to active persons (242 ± 22.3 vs. 186 ± 20.4 ms).

3.2.1.2. Local effects Responses were slower in switch than non-switch trials (1109 ± 30.2 vs. 976 ± 30.0 ms; F(1,38) = 34.5, p < .0001, η² = .476), documenting substantial switch costs. The effect of Group showed slower RTs in low active compared to active persons (1112 ± 39.5 vs. 974 ± 39.5 ms; F(1,38) = 6.1, p < .05, η² = .139). The interaction Task-Set Transition x Group was far from significance (F < 1). ISDs were larger in switch than non-switch trials (416 ± 13.3 vs. 362 ± 14.8 ms; F(1,38) = 14.8, p < .0001, η² = .281). The interaction Block x Group was not significant (F < 1). However, the variability of speed was generally larger in low active compared to physically active participants (418 ± 17.3 vs. 361 ± 17.3 ms; F(1,38) = 5.4, p < .05, η² = .125).

3.2.2. Error rates 3.2.2.1. Mixing effects The analysis of accuracy yielded enhanced error rates in the mixed than single task blocks (11.0 ± 1.6 % vs. 2.4 ± 0.4 %; F(1,38) = 33.0, p < .0001, η² = .465), suggesting 19

substantial mixing costs and a significant interaction Block x Group (F(1,38) = 5.1, p < .05, η² = .119) due to larger mixing costs on accuracy in the low active relative to active participants (11.9 ± 2.5 % vs. 5.1 ± 1.5 %, p < .05). The main effect of Group did not reach significance (F(1,38) = 2.2, p = .146, η² = .055).

3.2.2.2. Local effects Error rates were larger in switch than non-switch trials (19.0 ± 2.1 % vs. 11.0 ± 1.5 %; F(1,38) = 38.4, p < .0001, η² = .503) and larger in low active compared to active seniors (19.7 ± 2.5 % vs. 10.3 ± 2.5 %; F(1,38) = 7.2, p < .01, η² = .158). Moreover, the interaction Task-Set Transition x Group was also significant (F(1,38) = 6.8, p < .05, η² = .153), suggesting larger local costs in accuracy in low active than physically active participants (11.5 ± 2.3 % vs. 4.6 ± 1.1 %, p < .05).

3.3. ERP data In Figures 3 and 4 ERP-waveforms after (dummy-) cues are presented at Fz, Cz, and Pz for the physically active and low active participants as a function of mixing (Figure 3) and local effects (Figure 4). The corresponding target-locked ERPs are shown in Figures 5 and 6. The mean amplitudes of the target-locked P2, N2 and P3b are plotted in Figure 7.

3.3.1. Cue-locked In the preparation interval the relationships between physical activity and P2, N2, P3b and CNV were analysed.

3.3.1.1. Mixing effects ------------------------------20

Insert Fig. 3 about here. --------------------------------

The ERP analysis revealed an amplitude effect of Block on the P2 (F(1,38) = 11.8, p < .001, η² = .237), indicating a larger P2 in the single than the mixed block (1.9 ± 0.3 vs. 0.7 ± 0.3 µV). No effect of Group or interaction Block x Group was found (all F’s < 1). The P2 latency did not reveal any effects or interactions (all F’s < 1). Similarly, no effects were found regarding the N2 amplitude or latency (all F’s < 1). The P3b was smaller in the mixed than the single task block (0.4 ± 0.3 vs. 1.7 ± 0.2 µV, F(1,38) = 18.9, p < .0001, η² = .332). No further effects or interactions were found regarding the P3b amplitude or latency (all F’s < 1). The subsequent CNV did not reveal any significant effect of Block (-0.2 ± 0.4 vs. -1.2 ± 0.5 µV, F(1,38) = 3.0, p = .09, η² = .074) nor an interaction Block x Group (F < 1). However, there was an effect of Group (0.1 ± 0.5 vs. -1.5 ± 0.5 µV, F(1,38) = 4.5, p < .05, η² = .107), suggesting a less negative frontal CNV in the physically active than low active participants.

3.3.1.2. Local effects ------------------------------Insert Fig. 4 about here. --------------------------------

Task-Set Transition or Group showed no effects or interactions on P2 amplitude. However, the P2 peaked earlier in physically active than inactive seniors (209 ± 5.4 vs. 230 ± 5.4 ms; F(1,38) = 7.8, p < .01, η² = .170). Regarding N2, P3b and CNV no effects or interactions were found (all F’s < 1). 21

3.3.2. Target-locked 3.3.1.1. Mixing effects Figure 5 shows target-locked ERPs as a function of type of Block and Group. The P2 amplitude revealed no effects or interactions. However, as found in the cuelocked data the P2 peaked earlier in the active than the low active seniors (197 ± 4.1 vs. 215 ± 4.1 ms; F(1,38) = 9.3, p < .005, η² = .197). The N2 was considerably larger in the single than in the mixed block (-2.5 ± 0.5 vs. 0.4 ± 0.3 µV; F(1,38) = 55.8, p < .0001, η² = .595). Importantly, the N2 was generally larger in the physically active compared to the low active group (-2.3 ± 0.6 vs. 0.2 ± 0.6 µV; F(1,38) = 9.7, p < .005, η² = .203). The interaction Block x Group did not reach significance (F(1,38) = 2.5, p = .121, η² = .062). No effects or interactions were found for the N2 latency (all F’s < 1). The P3b was smaller in the mixed than in the single task block (4.8 ± 0.4 vs. 7.5 ± 0.5 µV; F(1,38) = 96.8, p < . 0001, η² = .718). No further effects or interactions on P3b were obtained (all F’s < 1). ------------------------------Insert Fig. 5 about here. --------------------------------

3.3.1.2. Local effects Figure 6 presents the target-locked ERPs in the mixed block as a function of Task-Set Transition and Group. The P2 latency peaked again earlier in active than low active participants (200 ± 4.6 vs. 213 ± 4.6 ms; F(1,38) = 4.3, p < . 05, η² = .102). No effects or interactions on P2 amplitude occurred (all F’s < 1). 22

The N2 amplitude was again larger in the physically active compared to low active seniors (-0.8 ± 0.5 vs. 1.1 ± 0.5 µV; F(1,38) = 8.6, p < . 005, η² = .185). N2 latency showed a trend to be delayed in switch vs. non-switch trials (333 ± 8.2 vs. 319 ± 7.0 ms; F(1,38) = 3.9, p = . 054, η² = .094). ------------------------------Insert Fig. 6 about here. -------------------------------The P3b revealed an interaction Task-Set Transition x Group (F(1,38) = 6.2, p < . 05, η² = .140), suggesting lower amplitude in switch vs. non-switch trials in the active group (3.8 ± 0.5 vs. 4.7 ± 0.6 µV; (1,19) = 7.8, p < . 05, η² = .235) and no difference in the low active group (5.3 ± 0.6 vs. 5.5 ± 0.5 µV; F < 1). No further effects or interactions occurred (all F’s < 1). ------------------------------Insert Fig. 7 about here. --------------------------------

In sum, the result of the target-locked data show an earlier P2, a larger N2 and a typical P3b pattern, i.e. a smaller P3b in switch than non-switch trials (e.g. Barceló et al., 2000; Gajewski & Falkenstein, 2011; Lorist et al., 2000; Rushworth et al., 2002), in the physically active compared to the low active seniors.

3.4. Correlations between behavioural and physical activity data In order to assess the relationship between the level of physical activity and behavioral performance a correlational analysis was conducted for the relationship between the power-to-weight ratio or self-reported activity level on the one hand and mixing or local costs in speed and accuracy on the other. We found negative relationships between 23

self-reported activity assessed by the Activity Questionnaire and mixing costs in accuracy (r = -0.44; p < 0.005), as well as local costs in accuracy (r = -0.52; p < 0.001).

4. Discussion 4.1 Behavioral data The aim of the present study was to investigate the association between lifelong physical activity and switching ability and its underlying neural mechanisms in older individuals. The active and low active groups were matched regarding gender, age, health status, medication and educational level. A number of neuropsychological tests assessing basic cognitive functions like speed of processing, attention or memory did not reveal any differences between groups apart from the Stroop task which revealed larger interference costs in low active compared to the active group. The results of the memory based switch task corroborated the relationship between compromised executive functions and lack of regular physical activity. This was evident in generally longer reaction times, a higher variability of speed and lower accuracy in low active participants, corroborating a number of previous studies. Moreover, we observed increased mixing costs in this group in all three behavioral parameters. Significant group differences in local costs were observed in accuracy only. In other words, the behavioral findings point to a relationship between regular physical activity and executive functions regarding different facets of task switching.

4.2 ERP data Electrophysiological data help to elucidate correlates of these differences and to shed more light on the underlying functional mechanisms. In the task preparation period we found a shorter P2 latency in the active compared to the low active group, suggesting 24

faster retrieval of the relevant task-sets. Moreover, the frontal CNV was lower in the active group which suggests lower effort in respect to cognitive control of the preparing task in physically active elderly, which is in line with previous research (Hillman et al., 2002; Kamijo et al., 2010). Thus, there is obviously a general tendency of lower frontal control in anticipation of a target in physically fit participants. Nevertheless, as the most consistent group differences were found after the target onset, it is plausible to assume that the group difference observed in the behavioral data was primarily due to processing of targets and response selection mechanisms. Therefore, we will consider target processing in more detail. The most obvious differences between our physically active and low active participants were found in the shorter latency of the P2 and a more pronounced N2 in active seniors, regardless of block and condition. As already mentioned in the introduction, the P2 probably reflects recall of action related instructions like S-R mapping (Adrover-Roig & Barceló, 2008; Finke et al. 2011; Kieffaber & Hetrick, 2005; Schapkin et al., 2014; van Elk et al., 2010), whereas the N2 seems to reflect response selection, i.e. the application of the retrieved S-R set (Gajewski et al. 2010a). Therefore, a smaller N2 in the grand average may indicate larger N2 latency jitter that leads in turn to increased variability of response. This supports also the pattern of larger N2 in single task blocks in which response selection is not compromised by interference due to overlapping S-R sets which resulted in faster and less variable responses (e.g. Gajewski & Falkenstein, 2012; Kray et al., 2005; Karayanidis et al., 2011; Schapkin et al., 2014). In difficult tasks with overlapping S-R mappings like in mixing switch block the N2 is delayed and blurred, suggesting a less synchronized implementation of S-R sets, which in turn enhances the reaction time variability (Gajewski et al., 2011; Gajewski & Falkenstein, 2012). This “system noise” may also reduce the probability to select the correct response. In agreement with this, we found 25

longer RTs, larger ISDs and lower accuracy paralleled with smaller N2 in low active compared to physically active participants. Contrary to our expectations the amplitude of the P3b was not generally larger in physically active compared to low active individuals. In addition, P3b amplitude was reduced in mixed relative to single task blocks regardless of group, supporting the notion of an inverse relationship between P3b and task difficulty (Kok, 2001). In addition, the P3b revealed an interesting group difference concerning task set transition in the mixed block: whereas no difference between non-switch and switch trials was found in the low active individuals, the amplitude was lower in task switch vs. non-switch trials in the physically active elderly, which is the typical pattern observed in young subjects (Lorist et al., 2000; Rushworth et al., 2002; Karayanidis et al., 2011). Hence, the active seniors show a pattern which is known from young subjects whereas the low active seniors do not show this pattern. Also, contrary to our prediction the P3b in the mixed block appears generally reduced in the active compared to the low active seniors. As apparent in Figures 5 and 6 this reduction is largest at Cz, where the N2 is also enhanced in the active seniors. This pattern supports the assumption of a negative shift in the active seniors which enhances the N2 and decreases the P3b. This shift may reflect an increased frontocentral processing in the active individuals with the aim to optimize performance. Thus, the general P3 amplitude pattern in active vs. inactive seniors is quite different: while the low active seniors show the typical flattening of P3 topography (similar amplitudes at Pz, Cz and Fz), the active seniors show a more youth-like pattern with P3 maximum at Pz and smaller (i.e. more negative) amplitudes at Cz and Fz.

4.3 Correlational findings

26

The subjectively reported level of physical activity was negatively correlated with mixing and local costs in accuracy, suggesting some dose-response relationship between the level of physical activity and these cognitive parameters.

4.4 Contribution to the literature The present study contributes to the relatively small body of research combining aging, physical activity and executive functions assessed by task switching using ERP measures. Our behavioral results are consistent with those of Hillman and colleagues (2006) who also observed lower mixing costs in physically active than low active participants. In contrast, local switch costs were reduced in Hillman’s active participants but did not differ between the groups in our study. A similar pattern of mixing and local costs like in the present study was found by Themanson et al. (2006b) and more recently in Dai et al. (2013). Regarding electrophysiological data, our findings contribute to the existing body of research by showing that a superior performance in a task switching paradigm in habitually physically active seniors is related to shorter P2 latencies and larger amplitudes of the N2 or rather a prolonged frontocentral negativity. A similar pattern was found in the same sample using a Stroop paradigm (unpublished data). We interpreted this pattern in terms of faster retrieval of S-R units (P2) and an enhanced cortical activity allocated to task relevant response processing (N2 and N450 in the Stroop paradigm). On the first glance, it may be surprising that enhanced brain activity is associated with enhanced behavioral performance as it would be plausible to assume fewer resources are needed to maintain a reasonable level of performance as suggested in some previous reports. Our results and some other studies support this notion by showing smaller preparatory activity as reflected in the CNV in physically fit than less fit 27

individuals (Hillman et al., 2002; Kamijo et al., 2010). Nevertheless, several ERP and fMRI studies indicate generally enhanced brain activity paralleled with enhanced performance in physically active individuals (Dai et al., 2014; Hillman et al., 2006; Gomez-Pinilla & Hillman, 2013; Prakash et al., 2011, see also Hayes et al., 2013 and Voelcker-Rehage & Niemann, 2013 for reviews of fMRI literature). It may be possible that physical fitness affects preparatory and target-related brain activity in opposite directions.

4.5 Limitations There are also a number of limitations which have to be acknowledged. The main disadvantage of cross-sectional studies is the inability to estimate the causal direction of associations. The second important limitation is the well known disadvantage of cross-sectional designs which makes impossible to control potential confounding variables, which may influence task performance and ERPs. We tried to avoid this by an optimal matching of the groups. The participants were matched as well as possible regarding a number of critical variables like age, education or health status. Nevertheless, small samples in randomized controlled studies may also led to asymmetric distribution of some variables. Third, the type, total duration and frequency of the physical activity were individually different. Also the ratio of aerobic, resistance, strength or coordinative training differed between the participants of the active group.

4.6 Future directions An open question is the time point when the active and low active groups begin to differ. In our previous randomized controlled study with training duration of 4months in seniors we found trends for reduction of mixing costs in speed and accuracy in the physical training group, but we did not observe similar ERP effects (though a tendency 28

to larger N2 in switch trials was observed, c.f. Gajewski & Falkenstein, 2012). We assume that the training duration was too short to substantially enhance brain activity. Future training studies may investigate ERPs at different time points or constitute trainings of different duration to extract the time point when the groups start to differ regarding behavior and ERPs. This may also help to elucidate the optimal training duration.

4.7 Conclusions In summary, this study provides evidence that habitual long-term physical activity positively affects executive control processes such as those measured by the memory based task-switching paradigm. The behavioral benefit in active persons was clearly evident in speed and accuracy domain. Electrophysiological data suggest that the origin of this benefit is localized in a faster retrieval of task-sets and more frontal resources allocated for the selection of the correct response. The findings suggest that habitual long-term physical activity enhances neuronal activity and integrity needed to perform complex cognitive tasks.

Acknowledgements This work was supported by a grant from the German Insurance Association (GDV, Gesamtverband der Deutschen Versicherungswirtschaft). We thank Ludger Blanke, Christiane Westedt, Brita Rietdorf, Claudia Wipking, Silke Joiko for excellent technical and organizational assistance. We wish also to thank Prof. Andreas Seeber for the idea to conduct this study, two anonymous reviewers for thoughtful comments on the earlier draft of this article and all participants.

29

References Adrover–Roig, D., & Barceló, F. (2010). Individual differences in aging and cognitive control modulate the neural indexes of context updating and maintenance during task switching. Cortex, 46, 434–450. Allport, A., Styles, E.A., and Hsieh, S., (1994). Shifting intentional set: exploring the dynamic control of tasks. In: Umiltà, C., Moscovitch, M. (Eds.), Attention and Performance XV. MIT Press, Cambridge MA, 421–452. Aschenbrenner, S., Tucha, O., & Lange, K. (2001). Der Regensburger Wortrflüssigkeits–Test. Göttingen: Hogrefe. Barceló, F., Muñoz–Céspedes, J. M., Pozo, M.A., & Rubia, F.J. (2000). Attentional set shifting modulates the target P3b response in the Wisconsin card sorting test. Neuropsychologia, 38, 1342–1355. Berchicci, M., Lucci, G., & Di Russo, F. (2013). Benefits of physical exercise on the aging brain: The role of the prefrontal cortex. J. Gerontol. A Biol. Sci. Med. Sci., 68, 1337–1341. Brickenkamp, R. (1994). Test d2, Aufmerksamkeits–Belastungs–Test (8th ed.). Goettingen: Hogrefe. Broadbent, D.E., Cooper, P.F., FitzGerald, P., & Parkes, K.R. (1982). The Cognitive Failures Questionnaire (CFQ) and its correlates. Br. J. Clin. Psychol., 21, 1–16. Brunia, C.H., & van Boxtel, G.J. (2001). Wait and see. Int. J. Psychophysiol 43, 59–75. Butler, K.M., & Weywadt, C. (2013). Age differences in voluntary task switching. Psychol. Aging, 28, 1024–1031. Cambell, P.T., Katzmarzyk, P.T., Malina, R.M., Rao, D.C., Pérusse, L., & Bouchard, C. (2001). Prediction of physical activity and physical work capacity (PWC 150) in young adulthood from childhood and adolescence with consideration of parental measures. Am. J. Human. Biol., 13, 190–196. 30

Cepeda, N.J., Kramer, A.F., & Gonzalez de Sather, J.C. (2001). Changes in executive control across the life span: Examination of task-switching performance. Dev. Psychol., 37, 715–730. Chang, M., Jonsson, P., Launer, L., et al. (2010). The effect of midlife physical activity on cognitive function among older adults: AGES-Reykjavik Study. J. Gerontol. A Bio. Sci. Med. Sci., 65, 1369–1374. Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences (second ed.). Lawrence Erlbaum Associates. Colcombe, S.J., Erickson, K.I., Raz, N., Webb, A.G., Cohen, N.J., Mc Auley, E., & Kramer, A.F. (2003). Aerobic fitness reduces brain tissue loss in aging humans. J. Gerontol. A Biol. Sci. Med. Sci., 58, 176–180. Colcombe, S.J., Erickson, K.I., Scalf, P.E., Kim, J.S., Prakash, R., McAuley, E., Elavsky, S., Marquez, D.X., Hu, L., & Kramer, A.F. (2006). Aerobic exercise training increases brain volume in aging humans. J. Gerontol. A Biol. Sci. Med. Sci., 61, 1166–1170. Colcombe, S., & Kramer, A.F. (2003). Fitness effects on the cognitive function of older adults: a meta–analytic study. Psychol. Sci., 14,125–130. Colcombe, S.J., Kramer, A.F., Erickson, K.I., Scalf, P., McAuley, E., Cohen, N.J., Webb, A., Jerome, G.J., Marquez, D.X., & Elavsky, S. (2004). Cardiovascular fitness, cortical plasticity, and aging. Proc. Natl. Acad. Sci. U.S.A., 101, 3316– 3321. Costa, P.T. Jr., & McCrae, R.R. (1992). Revised NEO personality inventory (NEO–PI– R) and NEO Five–Factor Investory (NEO–FFI) professional manual. Odessa, F.L, Psychological Assesment Ressources, Inc.

31

Dai, C.T., Chang, Y.K., Huang, C.J., & Hung, T.M. (2013). Exercise mode and executive function in older adults: an ERP study of task–switching. Brain Cogn., 83, 153–162. Diamond, A. (2013). Executive functions. Annu. Rev. Psychol., 64, 135–168. Di Russo, F., Taddei, F., Apnile, T., & Spinelli, D. (2006). Neural correlates of fast stimulus discrimination and response selection in top-level fencers. Neurosci. Lett., 408, 113–118. Donchin, E., & Coles, M.G.H. (1988). Is the P300 component a manifestation of context updating? Behav. Brain Sci., 11, 357–374. Draganski, B., & May, A. (2008). Training-induced structural changes in the adult human brain. Behav Brain Res., 192, 137–142. Dustman, R.E., Emmerson, R.Y., Ruhling, R.O., Shearer, D.E., Steinhaus, L.A., Johnson, S.C., Bonekat, H.W., & Shigeoka, J.W. (1990). Age and fitness effects on EEG, ERPs, visual sensitivity, and cognition. Neurobiol. Aging, 11, 193–200. Eppinger, B., Kray, J., Mecklinger, A., & John O. (2007). Age differences in task switching and response monitoring: Evidence from ERPs. Biol. Psychol., 75, 52– 67. Erickson, K.I., Raji, C.A., Lopez, O.L., Becker, J.T., Rosano, C., Newman, A.B., Gach, H.M., Thompson, P.M., Ho, A.J., & Kuller, L.H. (2010). Physical activity predicts gray matter volume in late adulthood: the Cardiovascular Health Study. Neurology, 75, 1415–1422. Erickson, K.I., Voss, M.W., Prakash, R.S., Basak, C., Szabo, A., Chaddock, L., et al. (2011). Exercise training increases size of hippocampus and improves memory. Proc. Natl. Acad. Sci. U.S.A., 108, 3017–3022.

32

Etnier, J.L., & Chang, Y.K. (2009). The effect of physical activity on executive function: A brief commentary on definitions, measurement issues, and the current state of the literature. J. Sport Exerc. Psychol., 31, 469–483. Falkenstein, M., Hoormann, J., Hohnsbein, J., & Kleinsorge, T. (2003) Short–term mobilization of processing resources is revealed in the event-related potential. Psychophysiology, 40, 914–923. Faul, F., Erdfelder, E., Lang, A.–G., & Buchner, A., (2007). G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behav. Res. Methods, 39, 175–191. Flöel, A., Ruscheweyh, R., Kruger, K., Willemer, C., Winter, B.,Volker, K., et al. (2010). Physical activity and memory functions: Are neurotrophins and cerebral gray matter volume the missing link? NeuroImage, 49, 2756–2763. Folstein, M. F., Folstein, S. E., & Mc Hugh, P. R. (1975). Mini–mental state. A practical method for grading the cognitive state of patients for the clinican. J. Psychiatr. Res., 12, 189–198. Folstein, J.R., & Van Petten, C. (2008). Influence of cognitive control and mismatch on the N2 component of the ERP: a review. Psychophysiology, 45, 152–70. Friedman, D., Nessler, D., Johnson, R., Ritter, W., & Bersick, M. (2007). Age–Related Changes in Executive Function: An Event-Related Potential (ERP) Investigation of Task–Switching. Aging Neuropsychol. Cogn., 15, 95–128. Friedman, D. (2008). The components of aging. In: Luck and Kappenman Oxford Handbook of Event–Related Potential Components. New York: Oxford University Press. Gajewski, P.D., & Falkenstein, M. (2011). Diversity of the P3 in the task-switching paradigm. Brain Res., 1411, 87–97. 33

Gajewski, P.D., & Falkenstein, M. (2012). Training-induced improvement of response selection and error detection in aging assessed by task switching: Effects of cognitive, physical and relaxation training. Front. Hum. Neurosci., 6, 130. Gajewski, P.D., Hengstler, J. Golka, K., Falkenstein, M., & Beste, C. (2011). The Met– allele of the BDNF Val66Met polymorphism enhances task switching in elderly. Neurobiol. Aging, 32, 2327.e7–e19. Gajewski, P.D., Stoerig, P., & Falkenstein, M. (2008). ERP—correlates of response selection in a response conflict paradigm. Brain Res., 1189, 127–134. Gajewski, P.D., Kleinsorge, T., & Falkenstein, M. (2010a). Electrophysiological correlates of residual switch costs. Cortex, 46, 1138–1148. Gajewski, P.D., Wild–Wall, N., Schapkin, S.A., Erdmann, U., Freude, U., & Falkenstein, M. (2010b). Effects of aging and job demands on cognitive flexibility assessed by task switching. Biol. Psychol., 85, 187–199. Gehring, W.J., Bryck, R.L., Jonides, J., Albin, R.L., & Badre, D. (2003). The mind's eye, looking inward? In search of executive control in internal attention shifting. Psychophysiology, 40, 572–585. Goffaux, P., Phillips, N.A., Sinai, M., & Pushkar, D. (2006). Behavioural and electrophysiological measures of task switching during single and mixed–task conditions. Biol. Psychol., 72, 278–290. Goffaux, P., Phillips, N.A., Sinai, M., & Pushkar, D. (2008). Neurophysiological measures of task-set switching: effects of working memory and aging. J. Gerontol. B Psychol. Sci. Soc. Sci., 6, 57–66. Gomez-Pinilla, F., & Hillman, C.H. (2013). The influence of exercise on cognitive abilities. Comprehensive Physiol., 3, 403–428. Gratton, G., Coles, M.G.H., & Donchin, E. (1983). A new method for off–line removal of ocular artifact. Electroencephalogr. Clin. Neurophysiol., 55, 468–484. 34

Guiney, H. & Machado, L. (2013). Benefits of regular aerobic exercise for executive functioning in healthy populations. Psychon. Bull. Rev., 20, 73–86. Hayes, S.M. Hayes, J.P., Cadden, M., & Verfaellie. M. (2013). A review of cardiorespiratory fitness–related neuroplasticity in the aging brain. Front Aging Neurosci., 5, 31. Hämmerer, D., Müller, V., & Li, S.C. (2014). Performance monitoring across the lifespan: Still maturing post-conflict regulation in children and declining task-set monitoring in older adults. Neurosci. Biobehav. Rev. 46, 105-123. Helmstaedter, C., & Durwen, H.F. (1990). VLMT: Verbaler Lern– und Merkfähigkeitstest. Ein praktikables und differenziertes Instrumentarium zur Prüfung der verbalen Gedächtnisleistungen. Schweiz. Arch. Neurol. Psychiatr. 141, 21–30. Hillman, C.H., Erickson, K.I., & Kramer, A.F. (2008). Be smart, exercise your heart: exercise effects on brain and cognition. Nat. Rev. Neurosci., 9, 58–65. Hillman, C.H., Kramer, A.F., Belopolsky, A.V., & Smith, D.P. (2006). A cross-sectional examination of age and physical activity on performance and event-related brain potentials in a task switching paradigm. Int. J. Psychophysiol., 59, 30–39. Hillman, C.H., Weiss, E.P., Hagberg, J.M., & Hatfield, B.D. (2002). The relationship of age and cardiovascular fitness to cognitive and motor processes. Psychophysiology, 39, 303–312. Höltke, V., & Jakob, E. (2002). Lüdenscheider Aktivitätsfragebogen zum Risikofaktor Bewegungsmangel. Lüdenscheid: Sportmedizin Hellersen. Hultsch, D.F., & MacDonald, S.W.S. (2004). Intraindividual variability in performance as a theoretical window onto cognitive aging. In: New Frontiers in Cognitive Aging, eds R.A. Dixon et al., Oxford University Press, Oxford, 65–88.

35

Hultsch, D.F., MacDonald, S.W.S., Hunter, M.A., Levy-Bencheton, J., & Strauss, E. (2000). Intraindividual variability in cognitive performance in older adults: Comparison of adults with mild dementia, adults with arthritis, and healthy adults. Neuropsychology, 14, 588–598. Jasper, H.H. (1958). The ten-twenty electrode system of the International Federation. Electroencephalogr. Clin. Neurophysiol., 10, 370–375. Jäger, A., & Althoff, K. (1994). Der WILDE–Intelligenz–Test. Göttingen: Hogrefe. Jost, K., Mayr, U., & Rösler, F. (2008). Is task switching nothing but cue priming? Evidence from ERPs. Cogn. Affect. Behav. Neurosci., 8, 74–84. Kamijo, K., O’Leary, K.C., Pontifex, M.B., Themanson, J.R., & Hillman, C.H. (2010). The relation of aerobic fitness to neuroelectric indices of cognitive and motor task preparation. Psychophysiology, 47, 814–821. Kamijo, K., & Takeda, Y. (2010). Regular physical activity improves executive function during task switching in young adults. Int. J. Psychophysiol., 75, 304–311. Karayanidis, F., Coltheart, M., Michie, P.T., & Murphy, K. (2003). Electrophysiological correlates of anticipatory and poststimulus components of task switching. Psychophysiology, 40, 329–348. Karayanidis, F., Whitson, L.R., Heathcote, A., & Michie, P.T. (2011). Variability in proactive and reactive cognitive control processes across the adult lifespan. Front. Psychol., 2, 318. Kieffaber, P.D., & Hetrick, W.P. (2005). Event–related potential correlates of task switching and switch costs. Psychophysiology, 42, 56–71. Kiesel, A., Steinhauser, M., Wendt, M., Falkenstein, M., Jost , K., Phillip, A., & Koch, I. (2010). Control and interference in task switching - A review. Psychol. Bull., 136, 849–874.

36

Kok. A. (2001). On the utility of P3 amplitude as a measure of processing capacity. Psychophysiology, 38, 557–577. Kramer, A.F., Hahn, S., & Gopher, D. (1999). Task coordination and aging: explorations of executive control processes in the task switching paradigm. Acta Psychol., 101, 339–378. Kramer, A.F., & Erickson, K.I. (2007a). Capitalizing on cortical plasticity: influence of physical activity on cognition and brain function. Trends Cogn. Sci., 11, 342–348. Kramer, A.F., & Erickson, K.I., (2007b). Effects of physical activity on cognition, wellbeing, and brain: Human interventions. Alzheimers Dement., 3, 45–51. Kray, J., & Lindenberger, U. (2000). Adult age differences in task switching. Psychol. Aging, 15, 126–147. Kray, J., Eber, J., & Lindenberger, U. (2004). Age differences in executive functioning across the lifespan: the role of verbalization in task preparation. Acta Psychol., 115,143–165. Kray, J. (2006). Task–set switching under cue-based versus memory-based switching conditions in younger and older adults. Brain Res., 1105, 83–92. Kray, J., Lucenet, J., & Blaye, A. (2010). Can older adults enhance task-switching performance by verbal self–instructions? The influence of working-memory load and early learning. Front Aging Neurosci., 147. doi: 10.3389/fnagi.2010.00147 Kray,J., Eppinger,B., & Mecklinger, A. (2005). Age differences in attentional control: an event–related potential approach. Psychophysiology, 42, 407–416. Lehrl, S. (1995). Mehrfach–Wortwahl–Test (MWT). Erlangen: Medizinische Verlagsgesellschaft. Lorist, M.M., Klein, M., Nieuwenhuis, S., De Jong, R., Mulder, G. & Meijman, T.F. (2000). Mental fatigue and task control: planning and preparation. Psychophysiology, 37, 614–625. 37

Los, S.A. (1996). On the origin of mixing costs: Exploring information processing in pure and mixed blocks of trials. Acta Psychol., 94, 145–188. Makeig, S., Westerfield, M., Jung, T.P., Covington, J., Townsend, J., Sejnowski, T.J., et al. (1999). Functionally independent components of the late positive event– related potential during visual spatial attention. J. Neurosci., 19, 2665–2680. Mayr, U. (2001). Age differences in the selection of mental sets: the role of inhibition, stimulus ambiguity, and response-set overlap. Psychol. Aging, 16, 96–109. Mayr, U., & Liebscher, T. (2001). Is there an age deficit in the selection of mental sets? Eur. J. Cognit. Psychol., 13, 47–69. Meiran, N., Gotler, A., & Perlman, A. (2001). Old age is associated with a pattern of relatively intact and relatively impaired task–set switching abilities. J. Gerontol. B. Psychol. Sci. Soc. Sci., 56, 88–102. Miyake, A., Friedman, N.P., Emerson, M.J., Witzki, A.H., & Howerter, A. (2000). The unity and diversity of executive functions and their contributions to complex frontal lobe tasks: A latent variable analysis. Cogn. Psychol., 41, 49–100. Moore, R.D., Wu, C.T., Pontifex, M.B., O’Leary, K.C., Scudder, M.R., Raine, L.B., & Hillman, C.H. (2013). Aerobic fitness and intra–individual variability of neurocognition in preadolescent children. Brain Cogn., 82, 43–57. Nicholson, R., Karayanidis, F., Poboka, D., Heathcote, A., & Michie, P.T. (2005). Electrophysiological correlates of anticipatory task-switching processes. Psychophysiology, 42, 540–554. Nicholson, R., Karayanidis, F., Davies, A., & Michie, P.T. (2006). Components of taskset reconfiguration: differential effects of ‘switch-to’ and ‘switch-away’ cues. Brain Res., 1121, 160–176. Polich, J. (1996). Meta–analysis of P300 normative aging studies. Psychophysiology, 33, 334–353. 38

Polich, J. (2007). Updating P300: an integrative theory of P3a and P3b. Clin. Neurophysiol., 118,128–48. Polich, J. & Lardon, M.T. (1997). P300 and long–term physical exercise. Electroencephalogr. Clin. Neurophysiol., 103, 493–498. Pontifex, M.B., Hillman, C.H., & Polich, J. (2009). Age, physical fitness, and attention: P3a and P3b. Psychophysiology, 46, 379–387. Potts, G.F. (2004). An ERP index of task relevance evaluation of visual stimuli. Brain Cogn. 56, 5–13. Prakash, R.S., Voss, M.W., Erickson, K.I., Lewis, J.M., Chaddock, L., Malkowski, E., et al. (2011). Cardiorespiratory fitness and attentional control in the aging brain. Front. Hum. Neurosci. 4, 229. Reitan, R.M. (1992). Trail Making Test: Manual for administration and scoring. Tucson, AZ: Reitan Neuropsychology Laboratory. Reimers, S. & Maylor, E.A. (2005). Task switching across the life span: Effects of age on general and specific switch costs. Develop. Psychol., 41, 661–671. Ritter, W., Simson, R., Vaughan, H.G.Jr., & Friedman, D. (1979). A brain event related to the making of a sensory discrimination. Science, 203, 1358–1361. Ritter W, Simson R, Vaughan H.G.Jr, & Macht M. (1982). Manipulation of event– related potential manifestations of information processing stages. Science, 218, 909–911. Rogers, R.D., & Monsell, S. (1995). The costs of a predictable switch between simple cognitive tasks. J. Exp. Psychol. Gen., 124, 207–231. Rushworth, M.F., Passingham, R.E., & Nobre, A.C. (2002). Components of switching intentional set. J. Cogn. Neurosci., 14, 1139–1150. Rushworth, M.F., Passingham, R.E., & Nobre, A.C. (2005). Components of attentional set-switching. Exp. Psychol., 52, 83–98. 39

Salthouse, T.A., Atkinson, T.M., & Berish, D.E. (2003). Executive functioning as a potential mediator of age–related cognitive decline in normal adults. J. Exp. Psychol. Gen., 132, 566–594. Schapkin, S.A., Gajewski, P.D., & Freude, G. (2014). Age differences in memory– based task switching with and without cues: An ERP study. J. Psychophysiol. 28. Scisco, J.L., Leynes, P.A., & Kang, J. (2008). Cardiovascular fitness and executive control during task–switching: an ERP study. Int. J. Psychophysiol., 69, 52–60. Smith, P.J., Blumenthal, J.A., Hoffman, B.M., Straumann, T.A., Welsch-Bohmer, K. Browndyke, J.N. et al. (2010). Aerobic exercise and neurocognitive performance: a meta-analytic review of randomized controlled trials. Psychosom. Med., 72, 239–252. Sterr, A., & Dean, P. (2008). Neural correlates of movement preparation in healthy ageing. Eur. J. Neurosci., 27, 254–260. Stroop, J.R. (1935). Studies of interference in serial verbal reactions. J. Exp. Psychol., 18, 643–662. Taddei, F., Bultrini, A., Spinelli, D., & Di Russo, F. (2012). Neural correlates of attentional and executive processing in middle–aged fencers. Med. Sci. Sports Exerc , 44, 1057–1066. Themanson, J.R., & Hillman, C.H. (2006a). Cardiorespiratory fitness and acute aerobic exercise effects on neuroelectric and behavioral measures of action monitoring. Neuroscience, 141, 757–767. Themanson, J.R., Hillman, C.H., & Curtin, J.J. (2006b). Age and physical activity influences on action monitoring during task switching. Neurobiol. Aging., 27, 1335–1345.

40

Themanson, J.R., Pontifex, M.B., & Hillman, C.H. (2008). Fitness and action monitoring: Evidence for improved cognitive flexibility in young adults. Neuroscience, 157, 319–328. van Boxtel, M.P., Paas, F.G., Houx, P.J., Adam, J.J., Teeken, J.C., & Jolles, J. (1997). Aerobic capacity and cognitive performance in a cross–sectional aging study. Med. Sci. Sports Exerc., 29, 1357–1365. Van Elk, M., Crajé, C. Beeren, M.E.G.V., Steenbergen, B. van Schie, H.T., & Bekkering. H. (2010). Neural evidence for impaired action selection in right hemiparetic cerebral palsy. Brain Res. 1349, 56–67. Van Veen, V., & Carter, C.S. (2002). The timing of action–monitoring processes in the anterior cingulate cortex. J. Cogn. Neurosci., 14, 593–602. Voelcker–Rehage, C., & Niemann, C. (2013). Structural and functional brain changes related to different types of physical activity across the life span. Neurosci. Biobehav. Rev., 37, 2268–2295. Voss, M.W., Prakash, R.S., Erickson, K.I., Basak, C., Chaddock, L., Kim, J.S., et al. (2010). Plasticity of brain networks in a randomized intervention trial of exercise training in older adults. Front Aging Neurosci, 2, 1–17. Voss, M.W., Nagamatsu, L.S., Liu–Ambrose, T., & Kramer, A.F. (2011). Exercise, brain, and cognition across the lifespan. J. Appl. Physiol., 111, 1505–1513. Voss, M.W., Erickson, K.I., Prakash, R.S., Chaddock, L., Kim, J.S., Alves, H., et al. (2013). Neurobiological markers of exercise-related brain plasticity in older adults. Brain Behav. Immun., 28, 90–99. Walter, W.G., Cooper, R., Aldridge, V.J., McCallum, W.C., & Winter, A.L. (1964). Contingent negative variation: an electric sign of sensorimotor association and expectancy in the human brain. Nature, 203, 380–384.

41

Wasylyshyn, C., Verhaeghen, P., & Sliwinski, M.J. (2011). Aging and task switching: a meta-analysis. Psychol. Aging, 26, 15–20. Wechsler, D., (1955). Manual for the Wechsler Adult Intelligence Scale. Psychological Corporation. New York. Weinstein, A.M., Voss, M.W., Prakash, R.S., Chaddock, L., Szabo, A., White, S.M. et al. (2012). The association between aerobic fitness and executive function is mediated by prefrontal cortex volume. Brain. Behav. Immun., 26, 811–819. Wendell, C.R., Gunstad, J., Waldstein, S,R., Wright, J.G., Ferrucci, L., & Zonderman, A.B. (2014). Cardiorespiratory fitness and accelerated cognitive decline with aging. J. Gerontol. A Biol. Sci. Med. Sci., 69, 455–462. West, R. (1996). An application of prefrontal cortex theory to cognitive aging. Psychol. Bull., 120, 272–292. West, R. (2004). The effects of aging on controlled attention and conflict processing in the Stroop task. J. Cogn. Neurosci., 16, 103–113. West, R., & Moore, K. (2005). Adjustments of cognitive control in younger and older adults. Cortex, 41, 570–581. West, R., & Travers, S. (2008). Differential effects of aging on processes underlying task switching. Brain Cogn., 68, 67–80. Wild-Wall, N., Falkenstein, M., & Gajewski, P.D. (2012). Neural correlates of changes in a visual search task due to cognitive training. Neural Plast. 2012, 529057. doi:10.1155/2012/529057. Wild-Wall, N., Hohnsbein, J., & Falkenstein, M. (2007). Effects of ageing on cognitive task preparation as reflected by event–related potentials. Clin. Neurophysiol., 118, 558–569.

42

Willemssen, R., Falkenstein, M., Schwarz, M., Müller, T., & Beste, C. (2011). Effects of aging, Parkinson’s disease, and dopaminergic medication on response selection and control. Neurobiol. Aging, 32, 327–335. Wu, C., Pontifex, M.B., Raine, L.B., Chaddock, L., Voss, M.W., Kramer, A.F., et al. (2011). Aerobic fitness and response variability in preadolescent children. Neuropsychology, 25, 333–341. Wylie, G.R., Javitt, D.C., & Foxe, J.J. (2003). Task switching: a high–density electrical mapping study. Neuroimage, 20, 2322–2342. Yeung, N., & Cohen, J.D. (2006). Impact of cognitive deficits on conflict monitoring. Predictable dissociations between the error related negativity and N2. Psychol. Sci., 17, 164–171.

43

Conflict of interest None of the authors had any financial or other conflict of interest regarding the methods and content of this paper.

Table captions Table 1 Sample characteristic and results of the Physical Work Capacity test (PWC-130) and self-reported physical activity level for the physically active and passive participants.

Number Age MMSE Level of education Weight [kg] Height [cm] BMI Activity level absolute power [Watt] relative power [Watt/kg]

low active n =20 73.6 (4.9) 28.7 (0.9) 3.6 (1.2) 85.1 (10.9) 176.0 (5.2) 27.4 (2.7) 1.6 (0.6) 109.2 (17.2) 1.3 (0.2)

active n =20 72.7 (4.3) 29.0 (0.8) 4.0 (0.6) 80.4 (9.2) 176.7 (6.2) 25.7 (2.6) 3.2 (0.7) 142.2 (20.7) 2.4 (0.3)

t(df)

p

t(38) =58 t(38) =1.23 t(38) =1.46 t(38) =1.46 t(38) =.38 t(38) =1.94 t(38) =8.43 t(38) =5.47 t(38) =5.74

.562 .229 .151 .153 .702 .060 .0001 .0001 .0001

Footnote for Table 1 Significance level was set at p < .05 Key: MMSE: Mini Mental State Examination; Mean Level of Education: No degree (1), Primary (2), Secondary general (3), Intermediate secondary (4), Gymnasium (5); BMI: Body Mass Index; Activity level: (1) low, (4) high; Absolute and relative power in the ergometry test PWC-130 (see section 2.2. for further information).

Table 2 Results of the neuropsychological testing for physically active and low active participants.

low

active

t(df)

p 44

active NEO-FFI neuroticism extraversion openess to experience agreeableness conscientiousness D2 total number of symbols (n) number omitted symbols (n) number confused symbols (n) Digit-Symbol-Test total number of symbols (n) Stroop word reading (sec.) color naming (sec.) interference list (sec.) interference index (sec.) Digit-span* forward (n) backward (n) word-fluency (n) MWT-B number total (n) IQ VLMT total score trials 1 to 5 (n) delayed recognition (n) Mental Rotation total number (n) number correct (n) TMT* TMT-A (sec.) TMT-B (sec.) CFQ total score (n)

t(38)=1.72 t(38)=0.65 t(38)=0.80 t(38)=3.11 t(38)=1.20

.094 .519 .429 .005 .221

395 (86.8) 363 (80.3) t(38)=1.20 18.1 20.6 t(38)=0.42 (19.7) (16.9)

.231

1.4 (0.6) 2.1 (0.5) 2.2 (0.5) 2.2 (0.5) 2.6 (0.6)

1.2 (0.4) 2.3 (0.4) 2.3 (0.4) 2.6 (0.3) 2.8 (0.4)

.671

2.9 (2.8)

3.3 (8.5)

t(38)=0.29

.846

41.2 (11.4)

45.3 (8.0)

t(38)=1.30

.192

14.0 (4.8) 23.0 (8.0) 48.0 (16.6) 24.9 (12.1)

14.3 (2.2) 22.2 (3.4)

t(38)=0.25 t(38)=0.44

.805 .664

40.2 (7.2)

t(38)=1.89

.067

18.0 (6.1)

t(38)=2.24

.031

7.7 (1.8) 6.0 (1.5) 41.4 (11.8)

7.5 (1.8) 6.2 (1.3) 48.1 (10.4)

t(38)=0.34 t(38)=0.43

.731 .663

t(38)=1.90

.065

32.7 (2.1) 123.1 (9.8)

33.5 (2.3) 127.7 (12.0)

t(38)=1.19

.242

t(38)=1.29

.205

36.4 (8.4) 12.0 (2.2)

38.3 (7.5) 12.1 (2.6)

t(38)=0.73 t(38)=0.13

.468 .646

7.7 (2.8) 6.4 (2.7)

7.8 (2.5) 6.0 (2.3)

t(38)=0.17 t(38)=0.49

.861 .621

41.5 (16.6) 105.0 (39.6)

39.3 (10.0) 92.8 (25.7)

t(38)=0.51

.613

t(38)=1.15

.257

29.2 (9.8)

31.7 (10.7)

t(38)=0.75

.454 45

Footnote for Table 2 Significance level was set at p < .05 Key: NEO-FFI: “Big Five” personality factors questionnaire; D2: Test of Attention; MWT-B: test of premorbid intelligence; VLMT: Verbal Learning and Memory Test; TMT: Trail Making Test; CFQ: Cognitive Failures Questionnaire (see section 2.3. for further information).

Figure captions Figure 1 Schematic illustration of a trial in the single task (top) and mixed task block. In contrast to the single task block, no explicit cue was presented in the memory based mixed block. The time structure in both blocks was exactly the same.

300 ms 1300 ms responseterminated

500 ms 500 ms

Figure 2

46

Mean reaction times (top) individual variability of speed expressed in standard deviations (middle) and error rates (bottom) with standard errors as a function of mixing effects (single task vs. non-switch) and local effects (non-switch vs. switch) for the physically active (grey bars) and low active (black bars) participants. The grey coloured areas indicate the time windows of ERP analyses.

RT [ms]

low active

active

1400 mixing effect

local effect

1200 1000 800 600 400

single task

non-switch

switch

ISD [ms] 500

local effect

mixing effect

400 300 200 100

single task

non-switch

switch

ERR [%] 30

local effect

mixing effect

25 20

15 10 5

single task

non-switch

switch

Figure 3 Grand average (dummy) cue-locked ERP-waveforms at Fz, Cz, and Pz for the physically active (red) and low active participants (black) as a function of mixing effects (single task: solid vs. non-switch trials from the mixing block: dashed).

47

The grey coloured areas indicate the time windows of ERP analyses.

-5 µV

Fz

mixing effects single task vs. non-switch trials

CNV

0 µV

P2

+5 µV Cz 0

500

1000

[ms]

N2

active single task active non-switch low active single task low active non-switch

Pz

P3b

Figure 4 Grand average (dummy) cue-locked ERP-waveforms at Fz, Cz, and Pz for the physically active (red) and low active participants (black) as a function of local effects (non-switch: dashed vs. switch: solid). The grey coloured areas indicate the time windows of ERP analyses.

48

-5 µV

Fz

local effects non-switch vs. switch trials

CNV

0 µV

P2

+5 µV Cz

0

500

1000

[ms]

N2

active non-switch

active switch low active non-switch low active switch Pz

P3b

Figure 5 Grand average target-locked ERP-waveforms at Fz, Cz, and Pz for the physically active (red) and low active participants (black) as a function of mixing effects (single task: solid vs. non-switch trials from the mixing block: dashed). The grey coloured areas indicate the time windows of ERP analyses.

49

-5 µV

Fz

mixing effects single task vs. non-switch trials

0 µV

+5 µV Cz

0

500

1000 [ms]

P2 active single task

N2

active non-switch

low active single task low active non-switch

Pz

P3b

Figure 6 Grand average target-locked ERP-waveforms at Fz, Cz, and Pz for the physically active (red) and low active participants (black) as a function of local effects (non-switch trials: dashed vs. switch trials: solid). The grey coloured areas indicate the time windows of ERP analyses.

50

-5 µV

Fz

local effects non-switch vs. switch trials

0 µV

+5 µV

0 Cz

500

P2

1000 [ms] active non-switch active switch low active non-switch

low active switch

N2

Pz

P3b

Figure 7 Mean amplitudes and standard errors of the target-locked P2 at Fz (top), N2 at Cz (middle) and P3b at Pz (bottom) for mixing and local effects in physically active (grey bars) and low active (black bars) participants.

51

low active

10 Fz [µV]

active

P2 8

mixing effect

local effect

6 4 2

single task 5

non-switch

Cz [µV]

3

switch

N2 mixing effect

local effect

1 -1 -3 -5 single task

non-switch

10 Pz [µV] 8

switch

P3b mixing effect

local effect

6 4 2

single task

non-switch

switch

Highlights Lifelong physical activity is related to lower mixing and switch costs in seniors. Physical activity is associated with lower frontal CNV during preparation. Physically fit group shows an earlier P2, and a larger N2 in the target-locked ERPs. Mixing and local costs are correlated with self-reported physical activity. Our data indicate an enhanced recall and implementation of task-sets in active seniors.

52