Effects of intermittent exercise during prolonged sitting on executive function, cerebrovascular, and psychological response: a randomized crossover trial

Abstract

Emerging evidence indicates that acute bouts of uninterrupted prolonged sitting decrease cerebral blood flow and impair executive function. Few studies have investigated the use of feasible sedentary behavior interruptions to attenuate these effects. This study aimed to investigate the effects of intermittent half-squat exercises during prolonged sitting on executive function. Twenty participants (45% women, 21 ± 1 yr) were randomized to sit for 3 h 1) without any interruptions (control) or 2) with 1 min half-squats every 20 min (exercise). Executive function was determined using the Color Word Stroop Test (CWST) and Trail Making Test-B (TMT-B). Subjective feelings of arousal and measures of fatigue, concentration, and motivation were evaluated. Internal carotid artery (ICA) blood flow was measured using Doppler ultrasound. There was a significant interaction effect for correct response times with the incongruent CWST (P < 0.01), which were 3.5% faster in the exercise and 4.2% slower in the control over 3 h of sitting. There was also a significant interaction effect for TMT-B completion times (P < 0.01), which were 10.0% faster in the exercise and 8.8% slower in the control. Exercise suppressed decreases in concentration with a significant interaction effect (−28.7% vs. −9.2% for control vs. exercise, P = 0.048) and increases in mental fatigue with a significant interaction effect (285% vs. 157% for control vs. exercise, P < 0.04). These changes may have been related to changes in ICA blood flow, which had a significant interaction effect (P = 0.087). These results suggest that a simple strategy like intermittent squat exercises could help to maintain executive function during prolonged sitting.

NEW & NOTEWORTHY We assessed executive function, cardiovascular, and cerebrovascular responses during 3-h prolonged sitting, with or without an exercise interruption (1 min squats every 20 min). Compared to uninterrupted sitting, exercise interruption suppressed sitting-induced reductions in cerebral blood flow and impairments in executive function. These results demonstrated the efficacy of a half-squat intervention for individuals seeking to preserve cognition during prolonged sitting, which may be useful in environments with limited resources such as the workplace.

INTRODUCTION

Engaging in chronic sedentary behavior is associated with the development of mild cognitive impairment (1–3). This association may be attributable to repeated exposure to acute bouts of prolonged sitting, one of the most prevalent forms of sedentary behavior, which has been shown to acutely impair cognitive function (4). For instance, there is evidence that executive function is compromised (5) and problem-solving errors tend to increase (6) during bouts of prolonged sitting. However, studies investigating the impact of interrupting prolonged sitting have produced mixed results. For example, executive function has been shown to improve (5), remain unchanged (7), or become impaired (4) during interrupted bouts of prolonged sitting.

One possible explanation to account for these discrepancies may relate to heterogenous study methods. For example, two studies assessed a 3-h sitting period (4, 7), while the other was 5 h (5). Similarly, the exercise interventions used were highly different, i.e., 10 calf raises every 10 min (4), a 3-min walk every 30 min (5), and 6 min high-intensity leg cycling after the first 1 h of the sitting period (7). The differences in these methods have made the development of physiological mechanisms underlying these changes in cognitive function difficult. Among the proposed mechanisms, a direct relationship between changes in total (8) or regional blood flow (4) to brain tissue has been suggested to be one of the most potent influential mediators of this trend (4, 8). There is additional evidence to support the role of indirect mechanisms, such as perfusion-related changes in glucose delivery (9) and cerebral oxygenation (4, 7) during prolonged sitting that may ultimately influence executive function. Previous studies have primarily used near-infrared spectroscopy (NIRS) to assess perfusion during prolonged sitting. However, NIRS may introduce confounding factors such as changes in skin blood flow at the forehead to measure cerebral oxygenation (10, 11). Using more global measurements, such as the assessment of volumetric cerebral blood flow using noninvasive methodologies such as ultrasound, may offer a clearer understanding of the effect of prolonged sitting on the cerebral circulation and any concomitant changes in executive function. Furthermore, although there is evidence for the role of physiological mechanisms influencing executive function, the role of psychological factors, such as mental arousal and mental fatigue, is even less clear. Indeed, previous research shows that mental fatigue (12, 13) or level of arousal (14) was associated with impaired cognitive performance. To address the above methodological shortcomings and determine the effect of sitting on cognition, comprehensive measurement including both physiological and psychological factors is required.

Based on the existing literature and following the recommendations of previous studies (4, 7), we aimed to investigate the effect of prolonged sitting (3 h), with or without interruptions in the form of half-squats every 20 min on cerebral blood flow and executive function. We hypothesized that squat interruptions could attenuate the decrease in executive function associated with prolonged sitting. To support our hypothesis, we also collected mechanistic physiological data, including measurements of volumetric cerebral blood flow using duplex ultrasound, along with related psychological variables such as mental fatigue and mental arousal.

METHODS

Study procedures were approved by the ethical committee of the Mount Fuji Research Institute and performed in accordance with the guidelines of the Declaration of Helsinki. This study was conducted according to the Consolidated Standards of Reporting Trials (CONSORT) guidelines (15). All study participants provided written informed consent before participation.

Participants

Data were collected as a part of a previous study (16). Twenty healthy young adults [45% women, mean age 21 (±1 standard deviation) years, mean BMI 21.6 (±1.6 standard deviation)·kg·(m²)⁻¹] were enrolled in the present study. Exclusion criteria based on individual self-report were current or expected pregnancy; regular engagement in moderate or vigorous physical activity (>120 min/wk); current smoking; history of cardiovascular or cerebrovascular diseases; and the use of medications that affect cardiovascular or cerebrovascular health. Results from one female participant in the previous study dataset were removed due to poor image quality of the internal carotid artery (ICA), while one male participant was newly recruited to increase the sample size back to 20 participants.

Experimental Design

This study comprised of a randomized crossover trial with two experimental conditions: prolonged (3 h) sitting without any interruptions (control; CON) and sitting (3 h) with 1 min of half-squat interventions every 20 min (exercise; EX, Fig. 1). The sitting period of 3 h is based on previous studies that investigated the effects of prolonged sitting on cognition (4, 7). The order of CON or EX was randomized. Participants visited the laboratory on three occasions: one familiarization visit and two experimental visits. The familiarization visit took place no more than 1 wk before the experimental visits. Participants were fully familiarized with the experimental procedures, the cognitive tasks, and the half-squat exercise protocol. Visits were separated by a minimum of 48 h to minimize carryover effects and a maximum of 4 days to minimize within-participant variation. All women had regular menstrual cycles and were studied on days 1–5 of their menstrual cycle to minimize the potential influence of hormonal variation (17, 18).

Figure 1.

Study procedure. BP, blood pressure; CWST, color word Stroop test; HR, heart rate; ICA, internal carotid artery; $\text{P}{\text{ET}}_{{\text{CO}}_2}$, partial pressure of end tidal carbon dioxide; TMT-B, trail making test-B; VAS, visual analogue scale.

Brachial blood pressure (BP), heart rate (HR), ICA blood flow, and partial pressure of end-tidal carbon dioxide ($\text{P}{\text{ET}}_{{\text{CO}}_2}$) were measured at 10, 60, 120, and 180 min into the sitting period. Participants’ arousal levels were evaluated presitting and postsitting. The participants completed a visual analog scale (VAS) for mental fatigue, concentration, and motivation, as well as a Color-Word Stroop Test (CWST) and a Trail-Making Test part B (TMT-B) presitting and postsitting. All study visits began between 07:00 and 11:00, and each participant arrived at the same time for both visits. Participants arrived fasted (excluding water consumption) for 12 h before visits and refrained from taking supplements on the morning of the experiment. In addition, they were asked to avoid strenuous physical activity and alcohol or caffeine consumption for 24 h before experimental visits. Upon arriving to the laboratory, participants were asked to void if possible. The experimental procedures were conducted in a quiet and environmentally controlled room. During the experimental conditions, participants were allowed to view a preselected, nonstimulatory video, but not allowed to consume any food or drink, nor go to the bathroom. Study personnel assisted with tasks such as laptop handling to control for participant muscle movement. Participants’ feet were placed on a nonslip mat to help avoid leg fidgeting. Study personnel monitored participants to ensure that they remained seated and did not fidget (Fig. 1).

Measurements

All measurements for cardiorespiratory variables, cerebral blood flow, cognitive function, and psychological indices were performed in a sitting position throughout the experimental period including pre, during, and post sitting.

Cardiorespiratory variables.

Participants’ height and body weight were measured at the beginning of each study visit. Then, participants rested in a supine position for at least 20 min while being fitted with cuffs on the upper arms, electrocardiogram electrodes, and a capnograph to measure end-tidal CO₂ (WEC-7301, Nihon Koden Co., Ltd., Tokyo, Japan). Furthermore, right carotid bifurcation was detected using Doppler ultrasound (set at 10.0 MHz) with a linear transducer (Logic-e; GE Healthcare, Tokyo, Japan). Systolic and diastolic BP were measured in both arms using oscillometry (Arterio Vision MS-3000, Osachi Co., Ltd., Yamanashi, Japan) at least twice for all participants, with a 1-min interval between repeated measurements. Measurements were repeated if the difference between the systolic or diastolic BP was >5 mmHg, in which case, a third measure was collected. The two closest values were then averaged (19). HR was measured using a portable HR monitor (Check-My-Heart, TRYTECH Co., Ltd., Tokyo, Japan) using the previously described methods (20, 21). Using these values, mean arterial pressure (MAP) was calculated as MAP = (systolic BP − diastolic BP) / 3 + diastolic BP. HR variability (HRV) was assessed by transferring the electrocardiogram recordings to a computer, and the data were analyzed for every 5 min of the electrocardiogram signal automatically using HRV analysis software. Time domain HRV was calculated using the standard deviation of normal-to-normal intervals (SDNN) and the root-mean-square of successive differences in the R-R interval (RMSSD). SDNN is considered an estimate of overall HRV, and RMSSD is an index of short-term components of HRV, which is mainly mediated by parasympathetic nerve activity (21, 22). In the frequency domain, the extent of very low-frequency oscillations (0.0033–0.04 Hz), low-frequency oscillations (LF: 0.04–0.15 Hz), and high-frequency oscillations (HF: 0.15–0.4 Hz) were quantified using a fast Fourier transformation (20–22). HF power and LF/HF ratio predominantly indicate the influence of both parasympathetic and sympathetic tones (18–20). $\text{P}{\text{ET}}_{{\text{CO}}_2}$ was measured using a pocket CO₂ monitor (WEC-7301, Nihon Kohden, Tokyo, Japan), and average values of 5 min were taken at each measurement timepoint.

Cerebral blood flow.

Right ICA measurements were performed 1.0– 1.5 cm distal to the carotid bifurcation using a Doppler ultrasound (set at 10.0 MHz) and a linear transducer (GE Logic-E, Healthcare, Tokyo, Japan). The reason for choosing this vessel is that 75% of global cerebral blood flow is supplied via the ICA (23–25), and a recent study demonstrated that ICA flow is associated with executive function in patients with heart failure though not reported in healthy individuals (26). For each measurement, the ICA blood flow was averaged over a period of 2 min at ∼5–10 min, 55–60 min, 115–120 min, and 175–180 min into the sitting period. The average ICA blood flow was calculated using the mean vessel diameter and flow velocity as previously described (21, 27). After obtaining a clear image of the vessel using the brightness mode, the mean vessel diameter was calculated as follows: mean diameter = (systolic diameter × 1/3) + (diastolic diameter × 2/3). The time-averaged mean flow velocity obtained using the pulse-wave mode was defined as the mean blood flow velocity (cm/s). Blood flow was calculated by multiplying the cross-sectional area × 60 (to convert mL/s to mL/min). Cerebrovascular conductance was calculated by dividing the ICA flow by MAP. During the measurements, we ensured that the probe position was stable, the insonation angle was constant (<60°), and the sample volume was positioned at the vessel’s center and adjusted to cover the vessel diameter’s width. ICA recordings were performed for 2 min at each time point using a commercial video capture device (AmCap, Microsoft, WA). The videos were analyzed offline using custom-designed edge detection and wall-tracking software (v. 2.0.1 No. S – 13037, Takei Kiki Kogyo, Japan) (21).

Cognitive function and psychological indices.

The CWST has been widely used in previous studies (4, 7) and was used to assess selective attention and conflict resolution (problem-solving). This test evaluates an executive assessment paradigm, measuring attention, response inhibition, interference, and behavioral conflict resolution (28). The three tasks (neutral, congruent, or incongruent) of the CWST were randomly administered on a computer screen with a pause of a few seconds between each trial. The CWST was a custom-made program using Excel VBA (29). During each CWST trial, 24 stimulus words were presented per task. Each task was randomly presented three times, resulting in nine trials presitting and postsitting.

Neutral stimuli are displayed in black text. Congruent stimuli are displayed as words that are the same as their color (i.e., the word “red” is displayed using red lettering), and incongruent stimuli are displayed as words where the text color and color word differ (i.e., the word “blue” is displayed using yellow lettering). The CWST presents the names of four colors (“Red,” Yellow,” “Green,” and “Blue”) displayed in incongruent colors (blue, green, red, and yellow, respectively) (modified using Excel VBA). We prepared a portable color-labeled numeric keypad. The numbers 2, 4, 6, 8, and 5 were labeled red, yellow, green, blue, and black, respectively. Before the CWST, the participants placed their right index finger on the black key (No. 5). During the CWST, participants pressed the color-labeled key (No. 2, 4, 6, or 8) that corresponded to the color of the stimulus word using only their index finger; they were asked to work as accurately and quickly as possible (29). The response accuracy and mean correct response time for each stimulus were used as indices of executive performance (29, 30).

The TMT-B (25), which is a valid and reliable test of executive function, was used in accordance with previous studies (9). The participants traced a line on paper that alternated between circled numbers and letters (i.e., 1, A, 2, B, 3, C…, totaling 25 circled numbers or letters). The time to completion was recorded, and the test was repeated three times and averaged (31); shorter times indicated greater cognitive flexibility (32). The test provides information regarding visual search speed, scanning speed, processing speed, mental flexibility, and executive function via assessment of set-shifting ability. The TMT-B reportedly has acceptable reliability (intraclass correlation coefficient: 0.85) in healthy adults aged from 18 to 84 yr (33).

To assess subjective psychological state, participants were given brief questionnaires immediately before completing cognitive tasks. The Felt Arousal Scale was used to assess mental arousal levels, ranging from 1 (low arousal) to 6 (high arousal) (34, 35). The assessment of mental fatigue, concentration, and motivation consisted of three sequential questions respective to each outcome. Each question was accompanied by a VAS (14, 36). Each VAS was labeled from 0 mm (not at all) to 100 mm (extremely). The participants drew lines to indicate their responses. In these tests, higher scores indicate increased arousal levels, greater mental fatigue, higher concentration, or greater motivation respectively.

Half-Squat Intervention

During the sitting period, participants in the EX condition performed 1 min of half-squats with calf raises between repetitions every 20 min at a rate of one repetition every 4 s (15 reps/min) which was tracked with a metronome. To perform the half-squats, the participants crossed their arms across their chests and bent their knees to 90°. To avoid isometric muscle contraction, the participants performed the half-squats without stopping when their knees were fully bent. This strategy has been shown in other studies to be a highly feasible, low-intensity exercise interruption across populations with beneficial effects on cardiometabolic disease risk factors (37, 38). Furthermore, this intervention does not have exercise resource limitations, meaning it can be performed more frequently throughout the workday without being limited by equipment availability.

Sample size.

To estimate the sample size required to detect the smallest detrimental (or beneficial) effect in a crossover study (39) with a Type I error rate of 0.05 and 80% power, ∼18 participants were required to detect a small effect change (0.2) in cerebral blood flow based on our previous study (31).

Statistics

The normality of the data was examined using Bartlett and Levene tests. If tests for equality of variance failed, the data were logarithmically transformed to create normality (19). Logarithmic transformation was conducted for RMSSD, HF, and LF/HF ratio data. Statistical analyses were performed using Jamovi 2.2.5, an open-source statistical program based in R coding. Only participants with complete data on primary outcomes were included in the analyses. The corresponding author had complete access to the study’s data and was responsible for the integrity of the dataset and data analyses. The effects of time (pre vs. post) and condition (CON vs. EX) were analyzed using linear mixed models, with the participant (intercept) specified as a random effect and time (slope) and condition specified as fixed effects. A statistical threshold of P = 0.10 was used to evaluate time × condition interaction effects to account for reduced statistical power (40), whereas a threshold of P = 0.05 was used for the evaluation of the main effects of time or condition. The effect size (ES) was calculated as Cohen’s d by dividing β by SD. ES was defined as trivial (<0.2), small (0.2), moderate (0.5), or large (0.8) (39).

RESULTS

Twenty participants were recruited, all of whom successfully completed both experimental trials with no missing data.

Cardiorespiratory Variables

Table 1 shows the cardiorespiratory variables during the 3-h sitting time between the CON and EX conditions. There was a significant interaction effect for HR (P = 0.074), with an average increase of 8.7% in the control condition and 5.1% in the exercise condition [Δ = −2.50 beats/min, 90% confidence interval (CI) = −4.76 to −0.25, ES= 0.78]. There were no significant time × condition interaction effects for any other cardiorespiratory variables, nor significant main effects for MAP. There were no significant interaction effects for any HRV outcome, but there were significant main effects for time for all HRV outcomes. SDNN, RMSSD, and HF decreased by −20.6% (Δ = −15.9 ms, 95% CI = −21.9 to −9.90, ES = −0.41 for SDNN), −13.0% [Δ = −0.21 ms, 95% CI = −0.24 to −0.18, ES = −0.96 for log (RMSSD)], and −14.9% [Δ =−0.36 ms, 95% CI = −0.44 to −0.30, ES= −0.85 for log (HF)], respectively. Log [LF/HF] increased by 110.3% (Δ = 0.15, 95% CI = 0.17 to 0.32, ES = 0.52). There were significant main effects of both time and condition on $\text{P}{\text{ET}}_{{\text{CO}}_2}$, which decreased from 0 to 180 min in both the CON and EX conditions (Δ = −1.0 mmHg, 95%CI = −1.60 to −0.38, ES = 0.50).

Table 1.

Central hemodynamics responses to uninterrupted sitting and to intermittent squat exercise

	10 Min	60 Min	120 Min	180 Min	Interaction	Time	Condition
MAP, mmHg
CON	77.8 ± 5.2	79.5 ± 6.4	80.0 ± 6.5	78.7 ± 7.3	β = 0.32	β = 1.05	β = 0.14
EX	78.3 ± 4.4	79.5 ± 5.9	79.5 ± 4.5	79.5 ± 5.1	P = 0.76	P = 0.048	P = 0.79
					ES = 0.07	ES = 0.43	ES = 0.06
HR, beats/min
CON	70.3 ± 7.5	75.7 ± 8.6	77.2 ± 7.9	76.3 ± 8.9	β = −2.50	β = 4.85	β = −1.23
EX	70.8 ± 8.2	75.7 ± 8.3	76.4 ± 8.1	74.4 ± 8.3	P = 0.074	P ≤ 0.001	P = 0.078
					ES = 0.41	ES = 2.04	ES = 0.03
$\text{P}{\text{ET}}_{{\text{CO}}_2}$, mmHg
CON	37.1 ± 2.3	36.4 ± 2.3	36.2 ± 2.6	35.3 ± 2.7	β = 0.83	β = −1.35	β = 0.42
EX	37.2 ± 2.5	36.9 ± 2.8	36.5 ± 2.5	36.3 ± 2.4	P = 0.02	P = <0.001	P = 0.02
					ES = 0.52	ES = 2.29	ES = 0.72
HR variability
SDNN, ms
CON	75.0 ± 35.3	68.1 ± 33.9	60.4 ± 21.1	56.7 ± 20.3	β = 4.78	β = −15.91	β = 4.41
EX	79.4 ± 34.3	69.4 ± 30.4	63.1 ± 29.8	65.9 ± 31.8	P = 0.46	P ≤ 0.001	P = 0.31
					ES = 0.165	ES = 0.215	ES = 0.78
Log (RMSSD), ms
CON	1.60 ± 0.20	1.49 ± 0.19	1.43 ± 0.19	1.38 ± 0.18	β = 0.03	β = −0.21	β = 0.02
EX	1.61 ± 0.19	1.51 ± 0.21	1.44 ± 0.18	1.42 ± 0.17	P = 0.31	P ≤ 0.001	P = 0.29
					ES = 0.23	ES = 2.50	ES = 0.21
Log (HF), ms2
CON	2.43 ± 0.35	2.19 ± 0.34	2.12 ± 0.32	2.04 ± 0.34	β = 0.01	β = −0.37	β = 0.01
EX	2.44 ± 0.35	2.24 ± 0.37	2.11 ± 0.33	2.08 ± 0.33	P = 0.89	P ≤ 0.001	P = 0.72
					ES = 0.03	ES = 1.85	ES = 0.06
Log (LF/HF), ratio
CON	0.23 ± 0.16	0.43 ± 0.20	0.45 ± 0.26	0.50 ± 0.24	β = −0.05	β = 0.25	β = −0.03
EX	0.22 ± 0.17	0.36 ± 0.25	0.43 ± 0.25	0.44 ± 0.23	P = 0.49	P ≤ 0.001	P = 0.45
					ES = 0.16	ES = 1.12	ES = −0.12

Values are means ± standard deviation (SD). n = 20 for each condition. CON, uninterrupted sitting control; ES, effect size; EX, intermittent squat exercise; HF, high-frequency; HR, heart rate; LF, low frequency; MAP, mean arterial pressure; $\text{P}{\text{ET}}_{{\text{CO}}_2}$, partial pressure of end-tidal carbon dioxide; RMSSD, root-mean-square of successive differences in R-R interval; SDNN, standard deviation of normal-to-normal intervals.

Executive Function

Regarding the response accuracy for the CWST, the accuracy of all trials (240 total trials) was >97%, and there was no significant interaction effect between time and condition, nor main effect of time or main effect condition (all P > 0.10). Figure 2 shows the correct response time of the CWST (Fig. 2, A–C) and the time to completion of the TMT-B (Fig. 2D) in response to prolonged sitting between the CON and EX conditions. There was a significant interaction effect for correct response time in the incongruent task (P = 0.008, Fig. 2A), with 4.2% slower response times for the CON condition and 3.5% faster response times in the EX condition (Δ = −40.43 ms, 90% CI = −64.84 to −16.02, ES = 0.61). There was a significant interaction effect for the correct response time in the congruent task (P = 0.017, Fig. 2B), with an average increase of 1.5% in the CON condition and an average decrease of 5.5% in the EX condition (Δ = −28.34 ms, 90% CI = −47.44 to −9.23, ES = 0.55). In contrast, the correct response time of the neutral task did not have a significant interaction effect (P = 0.124, Fig. 2C), nor significant main effects with time (P = 0.822) or condition (P = 0.079). A significant interaction effect was also found for TMT-B (P = 0.039, Fig. 2D), with 8.8% slower average completion times in the CON condition and 10.0% faster average completion times in the EX condition (Δ = −6.95 s, 90% CI = −10.89 to −3.02, ES = 0.44, Supplemental Table S1; all Supplemental Material is available at https://doi.org/10.6084/m9.figshare.24438955.v1).

Figure 2.

Changes in correct response time of CWST and time to completion of TMT-B. A: the changes in reverse-Stroop interference score at the pre- and post-CWST periods. Each panel illustrates the changes in incongruent (A), congruent (B), and neutral (C) tasks of CWST, and in TMT-B (D) in the presitting and postsitting periods between the control and exercise conditions. The values are shown as the means ± standard error (SE) and as individual data (dots). CWST, Color Word Stroop Test; ms, milliseconds; sec, seconds; TMT-B, Trail-Making Test Part B.

Psychological Variables

The psychological variables in response to prolonged sitting between the CON and EX conditions are shown in Fig. 3, A–D. There was a significant interaction effect for the level of mental arousal (P = 0.074, Fig. 3A), with mental arousal decreasing by 23% for CON and 1% in the EX condition (Δ = 0.70, 90% CI = 0.67–1.33, ES = 0.41). A significant interaction effect was also found in concentration (P = 0.048, Fig. 3B), which decreased by −28.7% for CON and by −9.2% for EX (Δ = 13.25 mm, 90% CI = 2.49– 24.01, ES = 0.45). Furthermore, there was a significant interaction effect for mental fatigue (P = 0.037, Fig. 3C), with mental fatigue increasing by 285% for CON and 157% for EX (Δ = −11.20 mm, 90% CI = −19.80 to −2.58, ES = 0.48). There was no significant interaction effect for motivation (P = 0.302), but there was a significant main effect with time (P < 0.001, Fig. 3D), with motivation decreasing by 19.4% on average across conditions (Δ = −14.57 mm, 95% CI = −20.74 to −8.41, ES = 1.03) for EX, with a nonsignificant main effect of condition (P = 0.087). Complete statistical data for psychological variables are shown in Supplemental Table S2.

Figure 3.

Changes in psychological variables for CWST and TMT-B. The panels illustrate changes in arousal level (A), ability to concentrate (B), mental fatigue (C), and motivation (D) in the presitting and postsitting periods between the control and exercise conditions. Arousal level was evaluated using a felt arousal scale. Mental fatigue, the ability to concentrate, and motivation were evaluated using a visual analog scale. The values are shown as the means ± SE and individual data (dots). CWST, Color Word Stroop Test; TMT-B, Trail-Making Test Part B.

Internal Carotid Artery

There was a significant interaction effect for ICA volumetric flow (P = 0.087), decreasing by 3.7% in the CON condition and increasing by 0.3% in the EX condition (Δ = 10.97 mL·min⁻¹, 90% CI = 0.46– 21.47, ES = 0.38, Table 2). Similarly, the ICA velocity had a significant time × interaction effect (P = 0.016), decreasing by 4.7% in the CON condition and 0.2% in the EX condition (Δ = 1.35 cm·s⁻¹, 90% CI = 0.24–2.41, ES = 0.47). The ICA diameter did not have a significant interaction effect (P = 0.705), nor were there main effects due to time (P = 0.318) without a significant effect by condition (P = 0.572). There were no significant interaction (P = 0.741) effects, nor main effects due to time (P = 0.090) or condition (P = 0.430) on the cerebrovascular conductance.

Table 2.

Internal carotid artery responses to without uninterrupted sitting and with intermittent squat exercise

	10 Min	60 Min	120 Min	180 Min	Interaction	Time	Condition
Blood flow, mL·min−1
CON	298 ± 49	302 ± 47	301 ± 42	287 ± 45	β = 10.96	β = −4.77	β = 5.45
EX	297 ± 52	305 ± 48	303 ± 49	298 ± 50	P = 0.09	P = 0.14	P = 0.09
					ES = 0.41	ES = −0.11	ES = 0.13
Blood flow velocity, cm·s−1
CON	28.1 ± 3.4	28.0 ± 3.4	27.8 ± 3.0	26.8 ± 2.3	β = 1.36	β = −0.65	β = 0.66
EX	28.0 ± 3.1	28.2 ± 3.3	28.1 ± 3.4	28.0 ± 3.5	P = 0.016	P = 0.02	P = 0.02
					ES = 0.05	ES = 0.07	ES = 0.07
Vessel diameter, mm
CON	4.75 ± 0.40	4.79 ± 0.42	4.80 ± 0.38	4.77 ± 0.40	β = −0.002	β = 0.002	β ≥ −0.01
EX	4.74 ± 0.39	4.80 ± 0.38	4.79 ± 0.40	4.76 ± 0.38	P = 0.63	P = 0.27	P = 0.63
					ES = 0.554	ES = 0.755	ES = 0.766
Vascular conductance, mL·min−1·mmHg−1
CON	3.84 ± 0.70	3.83 ± 0.72	3.80 ± 0.68	3.70 ± 0.73	β = 0.11	β = −0.10	β = 0.05
EX	3.81 ± 0.71	3.87 ± 0.70	3.83 ± 0.69	3.77 ± 0.71	P = 0.23	P = 0.03	P = 0.24
					ES = 0.27	ES = 0.62	ES = 0.34

Values are means ± SD. n = 20 for each condition. CON, uninterrupted sitting control; EX, intermittent squat exercise.

DISCUSSION

The major finding of the present study is that in response to 3-h sitting, half-squat interruptions could improve some aspects of executive function. This effect may be explained by preserved ICA blood flow with half-squats that also may explain maintained mental arousal, concentration, and feelings of fatigue. These data offer potential mechanistic insight into how cognition is preserved during prolonged sitting.

Limitations and Strengths

Before the discussion of the main findings, it is important to consider the strengths and limitations of this study, as they determine how the findings can be interpreted. A major strength of the study is the simultaneous measurement of physiological, cognitive, and psychological factors. This approach yields insight into potential physiological or psychological mechanisms underlying the changes in cognition seen during prolonged sitting. In addition, the role of psychological factors in affecting cognition during prolonged sitting is relatively unexplored in the current scientific literature. Lastly, the study used a highly generalizable exercise strategy (half-squats) that is both easily repeatable for future studies and easily implemented as an exercise strategy in any workplace setting.

However, the approach used in this study also had some limitations. First, this study used global, rather than local, measures of blood flow. Although previous literature has focused on the perfusion of local brain tissue using methods such as near-infrared spectroscopy (41–43), the authors of this study opted to use more global measures such as ICA blood flow. This approach was chosen to have a greater chance to detect changes in cerebral blood flow that co-occurs with changes in cognition. Furthermore, although far from definitive, there is evidence that changes in blood flow in large cerebral arteries correlate with changes in local cerebral perfusion as measured by near-infrared spectroscopy signals (11).

In this study, we only measured ICA flow, which constitutes 75% of the total cerebral blood flow (23–25). Although the lack of data regarding the vertebral artery, which supplies the remaining 25% of cerebral blood flow (23–25), may appear to be a weakness, a recent study found no changes in vertebral artery blood flow with prolonged sitting (31).

Second, participants were asked not to move their legs or fidget during the sitting bout except during half-squats, which was monitored by a research team member. Although this study design preserved the internal validity of our experiment by more effectively isolating the effect of the CON and EX conditions, this may limit the ecological validity of our results, as people are not restricted in this way in their daily lives. Similarly, participants were not allowed alcohol and caffeine 24 h before the experiment and we did not assess any effects of habitual consumption of these beverages on our outcomes. They were also not allowed to consume any food 12 h before the experiment. These study controls are commonly used in prolonged sitting studies (44, 45) but their effect on our study outcomes is unclear.

Third, we only tested one domain of cognitive function using two tests: the CWST and TMT-B. The CWST is a classic measure of prefrontal cortex function (46) that has been widely used to assess cognition (4, 7, 41–44), and the validity and the reliability of TMT-B have been confirmed (9). However, it is possible that other cognitive domains may be affected by prolonged sitting. Future studies may be able to assess other cognitive domains such as attention, memory, and decision-making.

Lastly, we purposely elected to recruit a small homogeneous group of young adults to avoid confounding factors such as age or training status effects. Therefore, the generalizability of our findings to older adults or adults with mild cognitive impairment may be limited. Also, although we recruited a similar number of male and female participants, we were underpowered to determine any sex-specific effects of the intervention. Further studies could evaluate any differences in the response of this intervention in older adults or adults with mild cognitive impairment, or any sex-specific effects to increase the generalizability of these results.

Comparison to Previous Literature

Our findings for cognition were somewhat in line with our initial hypothesis. Correct response times in both the incongruent and congruent tasks of the CWST and times to completion of TMT-B were impaired by acute prolonged sitting, which agrees with previous studies (4, 5). However, we did not find any changes in the neutral tasks of the CWST nor the accuracy for each task between the CON and EX conditions. Essentially, our results indicate that prolonged sitting affected both more difficult cognitive tasks (i.e., the incongruent task of the CWST and TMT-B) and simpler cognitive tasks (i.e., the congruent task of the CWST). We also found that these decreases were not observed with half-squat interruptions to sitting, where participants showed slightly improved reaction time and times to completion of the incongruent and congruent CWST task and TMT-B respectively relative to uninterrupted sitting. This contrasts with some prior literature, where local exercise [calf raises; (4) or exercise presitting (7)] did not improve executive function. It is possible that this discrepancy is due to the more frequent use of large muscle groups in the half-squat interruptions (quadriceps and posterior chain) compared to the use of small muscle groups [calves; (4) or only a single bout before sitting (7)]. This is supported by research indicating that both leg cycling (47–50) and whole body resistance exercise (51, 52), which similar to our study recruit a large number of muscle groups, showed an improvement in executive function across multiple populations.

We also observed an interaction effect for ICA flow between 10 and 180 min of sitting, with flow decreasing by 3.7% in the CON condition and increasing by 0.3% in the EX condition. These results are consistent with a prior study (53) and seem to show that although ICA flow is preserved with half-squats, it decreases with uninterrupted sitting. The decrease in flow seems to be driven by changes in ICA velocity, which also showed an interaction effect, with a 4.7% decrease in the CON condition and a 0.17% decrease in the EX condition. Meanwhile, ICA diameter showed a marginal increase (0.4%) across both the CON and EX condition with time. Decreased flow in the cerebral arteries during sitting may increase the risk for endothelial dysfunction (54), which is then prevented with half-squat interruptions.

Reduction in ICA flow in the CON condition might be partly explained by reductions in $\text{P}{\text{ET}}_{{\text{CO}}_2}$. $\text{P}{\text{ET}}_{{\text{CO}}_2}$ is proportionally associated with $\text{P}{\text{a}}_{\text{C}{\text{O}}_2}$ (55), which is critical in the control of cerebral blood flow (56). We found a reduction in $\text{P}{\text{ET}}_{{\text{CO}}_2}$ in the CON (−4.9%) condition compared to the EX condition (−2.4%) at 180 min of sitting. However, decreases in $\text{P}{\text{ET}}_{{\text{CO}}_2}$ in the EX condition (2.4% decrease) do not explain the maintenance of ICA flow (0.3% increase). Furthermore, blood gases primarily affect cerebral blood flow by altering cerebrovascular conductance, which was unchanged in this study (57). Future studies on the role of half-squats in affecting blood gas-related changes in cerebral blood flow may be able to clarify this differential effect.

Our results for psychological variables concur with the existing literature (13, 14). Although levels of arousal and ability to concentrate decreased and mental fatigue increased in both conditions, the magnitude of these changes was significantly less in the EX condition relative to the CON condition, suggesting that each of these factors was better maintained with half-squat interruptions. In agreement with this, a previous study found that intermittent walking during prolonged sitting could attenuate increases in subjective fatigue (58). Whether these improvements are related to hemodynamic changes in the cerebral vasculature or a direct effect of exercise is unclear. Previous literature indicates that exercise induces an improvement in executive function which is related to enhanced cerebral neural activity (41–43). Future research may be able to assess the mediating role, if any, of preserved hemodynamics in this relationship.

In the present study, HRV metrics did not show any differences between the CON and EX, suggesting that cardiac autonomic nervous activity may not have affected cerebral blood flow responses between conditions. However, HRV metrics did decrease across both conditions, indicative of a reduced influence of the parasympathetic nervous system. Furthermore, HR could be considered a measure of sympathetic nervous system activity and increased on average in both conditions, with an 8.7% increase in the CON condition and a 5.1% increase in the EX condition. Since the cerebral vasculature is richly innervated by sympathetic nerves, alterations in sympathetic nerve activity may affect the ICA flow in other contexts (59). Results for individual frequency measures (SDNN, RMSSD, HF) and LF/HF ratio indicate progressive reductions in parasympathetic nerve-mediated indices in both conditions. Our results indicate that prolonged sitting with and without interruption induced stressful conditions that reduced the influence of the parasympathetic nervous system and activated the sympathetic nervous system. Interestingly, HR increases were greater with uninterrupted sitting than with interrupted sitting.

Implications

The current findings suggest that half-squat exercises, a simple and highly feasible exercise strategy, can improve some aspects of executive function and suppress fatigue increases while maintaining ICA flow during acute bouts of prolonged sitting compared to uninterrupted sitting. Compared with other types of physical activity intervention such as walking, half-squats are practical as they can be performed at an individual’s desk and do not require specialized resources such as a treadmill or room to walk. Furthermore, resistance exercises, such as half-squats, are likely to be as effective as walking for the preservation of cardiovascular function with exposure to sitting (38). Lastly, the results of this provide associative support for our proposed mechanism, as changes in ICA blood velocity coincided with changes in the results of multiple cognitive tasks. Chronic reductions in brain blood flow, such as those that occur during uninterrupted sitting, may relate to increased dementia risk (60). Future studies can use our experimental model to test this longitudinally. Future studies may also be able to investigate changes in additional cognitive domains, such as attention, during prolonged sitting, and compare them to the results of this study.

Conclusions

Our results indicate a 3-h bout of uninterrupted prolonged sitting decreased ICA blood flow and decreased both accuracy and reaction time for difficult tests of cognition in the study sample (healthy untrained young men and women). We also found that half-squat exercises, performed for ∼1 min every 20 min during a 3-h bout of prolonged sitting and could maintain ICA blood flow and some aspects of cognition relative to uninterrupted sitting. Our half-squat intervention may be able to be used by individuals seeking to break up their sedentary behavior in an effort to preserve cognition during times, such as in the workplace.

DATA AVAILABILITY

Data will be made available upon reasonable request.

SUPPLEMENTAL DATA

Supplemental Tables S1 and S2: https://doi.org/10.6084/m9.figshare.24438955.v1.

GRANTS

This work was supported by the Japan Society for the Promotion of Science (Grant No. JP18K11012 to M.H.). K. Stone is supported by the Health and Care Research Wales funded National Cardiovascular Research Network, Wales.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M.H., Y.H., and L.S. conceived and designed research; M.H. and Y.H. performed experiments; M.H., A.P., Y.H., K.S., and L.S. analyzed data; M.H., A.P., Y.H., K.S., and L.S. interpreted results of experiments; M.H. and A.P. prepared figures; M.H., A.P., and L.S. drafted manuscript; M.H., A.P., Y.H., K.S., and L.S. edited and revised manuscript; M.H., A.P., Y.H., K.S., and L.S. approved final version of manuscript.