Procedures

Participants visited the laboratory five times. During their first visit, they filled out a questionnaire concerning physical activity status (International Physical Activity Questionnaire [27]), underwent a medical screening by a medical doctor, and performed an incremental cycle test to exhaustion. To minimise the learning effect, this test session was followed by two familiarisation sessions, taking place on two consecutive days. During these familiarisation sessions, the participants practised the cognitive test battery while sitting on a conventional chair (familiarisation session 1) and while cycling on the bike desk (familiarisation session 2). Furthermore, during both familiarisation sessions, participants were asked to cycle for 30 minutes, at 30% Wmax, in combination with their normal work activity. This way, the participants had the opportunity to become familiar with using the bike desk.

One week after familiarisation session 2, the participants returned to the lab twice, once to perform the cognitive test battery while sitting on a conventional chair, and once while cycling at 30% of their maximal external power (Wmax) on the bike desk. These two test moments were counterbalanced and were conducted one week apart at the same time of the day (between 9 and 12 am) to avoid influence of the circadian rhythm. Participants were not allowed to consume alcohol and to engage in vigorous physical activity 24 hours before the test moment. Furthermore, they were asked to consume a prescribed breakfast. Caffeine consumption on the test days was not allowed.

At the start of the test sessions, participants were fitted with a polar heart rate monitor and transmitter (Polar X-Trainer Plus, Polar Electro OY, Kempele, Finland) and the electroencephalography (EEG) cap (Acticap, Brain Products, Munich, Germany) was attached. This was followed by a practice Stroop test and typing performance trial. During these practice trials, computerised feedback about accuracy and speed was given. Subsequently, the participants performed a test battery consisting of the ‘Rey auditory verbal learning test’ (RAVLT), the ‘Stroop test’ with electroencephalography (EEG) recording, a typing test and the ‘Rosvold continuous performance test’ (RCPT) with EEG recording. These tests are described in detail below. The test battery took about 30 minutes. Participants continuously cycled, without any breaks.

The desk of the in height adjustable LifeSpan C3-DT5 Bike Desk was combined with an electrically braked cycle ergometer (Excalibur Lode, Groningen, the Netherlands). Power output was delivered in an isoinertial way (constant power), meaning that participants simply had to pedal at a self-selected pedalling rate without having to pay attention to hitting the target power output of 30% Wmax. Participants adjusted the cycle ergometer and the table until they were sitting in a comfortable posture. They had the possibility to re-adjust the cycle ergometer and the table after the practice Stroop and typing trial.

Age was self-reported. Height, body weight and fat percentage were measured to the nearest 0.1 cm, 0.1 kg and 0.1%, respectively. Peak oxygen uptake capacity (VO₂peak) was measured using an indirect calorimetry system (Metalyzer II^®, Cortex Biophysik, Leipzig, Germany) during an incremental cycle test to exhaustion. Participants started cycling at 50 Watts. Every three minutes, the load increased 25 Watts. The participants were asked to maintain a constant rhythm of 80 rpm. They were encouraged to exert themselves until volitional exhaustion. The decision to stop was based on signals of extreme fatigue and was confirmed by a heart rate that approximated the theoretical maximum heart rate (220-age) or a respiratory exchange ratio >1.10. VO₂peak was defined as the highest VO₂ attained over 30 sec. The maximal exercise test was performed on an electrically braked cycle ergometer (Excalibur Lode, Groningen, the Netherlands). Heart rate and rate of perceived exertion were assessed during the cognitive test batteries. Heart rate was recorded every 30 seconds using a polar heart rate monitor and transmitter (Polar X-Trainer Plus, Polar Electro OY, Kempele, Finland). Participants were asked to indicate their Rate of Perceived Exertion (RPE) on a modified 10-point Borg scale (Borg, 1982) at the end of both test batteries.

The transcription test was performed using ‘TypingMaster Pro’ (TypingMaster, Inc., Helsinki, Finland). Participants were provided with a different text of similar difficulty during each test moment. The order of both texts was counterbalanced. The participants were asked to transcribe as much of the text as possible within five minutes, while making as few mistakes as possible. Participants could not rely on spell check and could not go back in the typed text. They could only correct within the word they were typing. Typing speed and the number of mistakes were registered. Performance on the test was expressed in adjusted words per minute (AWPM) [14, 28].

The Dutch (native speech of the participants) version of the Rey auditory verbal learning test was used to assess short-term memory [19]. Fifteen words were five times read aloud by a trained staff member. Participants were each time asked to recall as many words as possible. After 20 minutes, at the end of the test battery, participants were asked to recall as many words as possible and to recognise the words of the list within a list of 30 read aloud words. The sum of recalled words of the five first trials, the amount of recalled words of the recall session and the amount of correctly and incorrectly recognised words during the recognition trial were used as outcome measures.

The Stroop test was programmed and performed on E-prime 2.0 software (Psychology Software Tools, Inc., Pittsburg, PA). The Stroop test was used to measure selective attention and response inhibition [29, 30]. The test consisted of three parts [16]. In the first part, measuring neutral reaction time, participants were demonstrated with X’s coloured in yellow, red, blue and green, and were asked to respond by pushing the corresponding button on a keyboard (AZERTY; F—left middle finger, V—left index finger, B—right index finger, H—right middle finger). In the second part, the words yellow, red, blue and green were shown in matching colours (congruent condition) and non-matching colours (incongruent condition). Participants were asked to push the button corresponding to the colour in which the words were displayed. In the third part, again the words yellow, red, blue and green were shown in matching colours (congruent condition) and non-matching colours (incongruent condition). This time, the participants were asked to push the button corresponding to the word displayed on the screen. The three parts were separated by a 30-sec rest period. Sixty stimuli were presented in the first part, 60 congruent and 60 incongruent in the second part, and 60 congruent and 60 incongruent in the third part. The interval response—stimulus onset was set at 500ms. The stimuli were displayed in the middle of the computer screen. The total time to perform the three parts of the test was approximately four minutes. Outcome measures were accuracy (%) and reaction time (ms).

The Rosvold continuous performance test (RCPT) was programmed and performed on E-prime 2.0. software (Psychology Software Tools, Inc., Pittsburg, PA). The RCPT was used to measure sustained attention [29, 31]. Over a time span of seven minutes, letters were continuously (every 1000 ms) presented to the participants. The participants were asked to push the space bar when an X appeared on the screen. Accuracy (%) and reaction time (ms) were assessed.

Continuous EEG was registered during the Stroop test and the RCPT using BrainVision Recorder (Brain Products GmbH, Munich, Germany). EEG data were derived from 32 active Ag/AgCl electrodes attached on the subjects’ head (Acticap, Brain Products, Munich, Germany) according to the ‘10-20 International System’ [32]. Electrode impedance was kept < 5kΩ throughout the experiment. Continuous data were recorded at a sampling rate of 500 Hz. To minimise sound artefacts, subjects were provided with earplugs.

ERP data were analysed using BrainVision Analyzer (Brain Products GmbH, Munich, Germany). Latency (ms) and amplitude (μV) were assessed for the N200, P300, N450 and conflict SP in response to the different stimuli during the Stroop test and for the N200 and P300 in response to the RCPT. The ERP data were processed as follows. Raw data were filtered (high pass: 0.1 Hz, low pass: 45 Hz and notch: 50 Hz; slope: 48dB/oct) with a Butterworth filter design and re-referenced to an average reference. Subsequently, artefacts were removed semi-automatically. Gradient, max-min, amplitude and low activity were set at 75 μV/ms, 150 μV/200ms, -100 μV, +100 μV and 0.5 μV/50ms respectively. Thereafter, the dataset was segmented (-200ms pre to 800ms post stimulus) into stimulus-locked epochs for correct trials. For each stimulus-locked epoch, artefacts were further removed using Independent Component Analysis (ICA) and inverse ICA. Baseline correction (using pre stimulus period -200 ms to 0 ms) was applied, the epochs were averaged, and N200, P300, N450 and SP were automatically detected. The number of trials contributing to the averages used in further analysis can be found in Table 1. The latency and amplitude for each ERP component were quantified using the mean amplitude and corresponding latency within a 150-300ms latency window for the N200, a 250-500ms latency window for the P300, a 450-550ms latency window for the N450 and a 600-800ms latency window for the conflict SP [18, 33–35]. Thereafter, the BrainVision Analyzer data were exported to IBM SPSS Statistics 22 for further analyses. The N200 and the N450 emerge fronto-centrally, while the P300 and the conflict SP emergs in temporal-parietal areas [18, 36, 37]. Therefore, we used the fronto-central region including Fp1, Fp2, F4, Fz, F3, F7, F8, FC1 and FC2 to analyse the N200 and the N450 and the temporal-parietal region including Pz, P3, P4, P7, P8, PO9 and PO10 to analyze the P300 and the conflict SP.