G Model
GAIPOS-4193; No. of Pages 8 Gait & Posture xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Gait & Posture journal homepage: www.elsevier.com/locate/gaitpost
Effect of dual tasking on postural responses to rapid lower limb movement while seated on an exercise ball P. Jones a, I. Sorinola b, P.H. Strutton a,* a The Nick Davey Laboratory, Human Performance Group, Division of Surgery, Department of Surgery and Cancer, Faculty of Medicine, Imperial College London, Charing Cross Hospital, London W6 8RF, UK b Division of Health and Social Care Research, School of Medicine, King’s College London, SE1 1UL, UK
A R T I C L E I N F O
A B S T R A C T
Article history: Received 13 March 2013 Received in revised form 3 April 2014 Accepted 8 April 2014
Postural adjustments are used by the central nervous system to pre-empt and correct perturbations in balance during voluntary body movements. Alteration in these responses is associated with a number of neuromuscular/musculoskeletal conditions. Attention has been identified as important in this system; performing a concurrent cognitive task has been suggested to reduce the efficacy of this postural control. The aim of this study was to examine the effect of concurrent cognitive tasking on anticipatory postural adjustments while sitting on an exercise ball with a view to help inform future rehabilitation programmes. Bilateral EMG activity was recorded from the external and internal obliques, rectus abdominis, erector spinae and the right rectus femoris of 20 healthy subjects (9 males) with mean (SD) age of 21.88 (0.86) years (range 21–24 years). A rapid hip flexion protocol was carried out under three conditions: no concurrent task, counting out loud up from one and completing a serial sevens task. The addition of the cognitive task delayed and reduced the EMG in the prime mover muscle but had little impact on the responses of the trunk muscles within the time frame of the anticipatory responses; suggestive of a decoupling of voluntary and postural control mechanisms. The results of this study suggest that perhaps the clinical effects of dual task may not be largely due to changes in anticipatory postural adjustments. However, it would be important to compare these results to those seen in older and functionally impaired individuals as this would be more representative of the typical population undertaking such rehabilitation programmes. ß 2014 Elsevier B.V. All rights reserved.
Keywords: Dual tasking Postural adjustments Trunk muscles Exercise ball Electromyography
1. Introduction Voluntary body movements generally displace the body’s centre of mass (COM) causing a disruption of posture [1,2]. In order to correct this shift in the COM, the central nervous system makes two overlapping phases of adjustments to trunk muscle activation. The first phase, termed anticipatory postural adjustments (APAs), begins before or is concurrent with the movement taking place; and aimed at minimising the postural effects of the movement [3]. These APAs are pre-planned and activated in a specific pattern by the CNS based on previous experience of the
* Corresponding author at: The Nick Davey Laboratory, Human Performance Group, Division of Surgery, Department of Surgery and Cancer, Faculty of Medicine, Imperial College London, Charing Cross Hospital, London W6 8RF, UK. Tel.: +44 (0) 203 313 8837; fax: +44 (0) 203 313 8835. E-mail address: [email protected] (P.H. Strutton).
effects of similar movements [4]. The second set of adjustments, compensatory postural adjustments (CPAs), counteract the actual perturbation in balance caused by the movement and are initiated by somatosensory feedback [5]. APAs are activated in the contexts of predictable perturbations such as self-generated movements; if the perturbations are unpredictable, APAs are often not observed [1]. CPAs, on the other hand, are required for both predictable and unpredictable perturbations and are initiated by somatosensory feedback. The mechanism of interaction between CPAs and APAs is not completely understood. However, there is evidence to suggest that stronger APAs reduce the need for large CPAs to preserve body equilibrium [1]. Deficits in postural adjustments [6] associated with limb movements, such as delayed activation and reduced magnitude of postural muscle activities have been described in a number of neuromuscular and musculoskeletal conditions [7,8], with consequences for functional abilities [9]. Therefore, strategies to improve APAs activations are a common construct in the
http://dx.doi.org/10.1016/j.gaitpost.2014.04.185 0966-6362/ß 2014 Elsevier B.V. All rights reserved.
Please cite this article in press as: Jones P, et al. Effect of dual tasking on postural responses to rapid lower limb movement while seated on an exercise ball. Gait Posture (2014), http://dx.doi.org/10.1016/j.gaitpost.2014.04.185
G Model
GAIPOS-4193; No. of Pages 8 2
P. Jones et al. / Gait & Posture xxx (2014) xxx–xxx
exercise programmes provided by physiotherapists. These programmes are often targeted at improving trunk activity and these can involve the use of exercise balls [10,11]. Although exercise balls are used extensively in motor rehabilitation the underlying mechanisms of any effects are still relatively unexplored. We have recently shown greater amplitudes of APAs and CPAs in trunk muscles occurring during lower limb movement while seated on an exercise ball compared to the same movement performed on a stable chair [12]. It was also suggested that there was a trend for early APA muscle activation while on the exercise ball compared to the chair [12]. These observations suggest that exercise balls might exert their influence on postural control system by modulation of the amplitudes and timing of APAs and CPAs. Postural control is often challenged during dual-tasking situations, when limb movements involved in activities of daily living tasks such as standing, walking and transfers are performed simultaneously with another task (i.e. cognitive or motor) [13]. To improve the ability to withstand these challenges, training under dual task conditions is often added to balance rehabilitation to facilitate better postural performance in these situations. Although improvements in postural control under dual task conditions [14] have been demonstrated with training, its effect on APAs and CPAs is still unclear. It is generally accepted that concurrent tasks and postural control either share available information processing capacity or the CNS organises their execution serially [15]. Therefore, if attentional resources are exceeded under dual tasking conditions, the performance of one or both tasks may become degraded. Previous studies have reported varying impacts of dual tasking on postural control. For example, increased stance sway was reported with concurrent performance of tasks with visuo-spatial components [16,17], or those that require verbal responses [18,19] or manipulation of surface and visual surround as in the Sensory Organisation Test [20,21]. In addition, it has also been found that the response EMG amplitude of postural muscles to perturbation was reduced under a dual task paradigm [22]. In contrast to these, other studies have found enhanced stance postural control [13,23] especially under dual task conditions with low cognitive demands. These observations suggest a varying impact on dual performance of cognitive tasks depending on the demands of both elements of the dual task paradigms. The effect dual tasking has on APAs and CPAs is currently unknown especially when combined with an exercise ball. Consequently, we sought to investigate the effect of dual tasking on postural adjustments to lower limb movements performed when seated on an exercise ball. We hypothesised that performing a simultaneous cognitive task will alter the magnitude and delay the timing of APAs and CPAs in trunk muscles while seated on an exercise ball.
2. Methods 2.1. Subjects With local ethical approval and written informed consent 20 healthy participants were recruited (9 males), with mean (SD) age of 21.88 (0.86) years (range 21–24 years), height 175.71 (9.15) cm and weight 69.71 (10.68) kg. Hand dominance was assessed using the Edinburgh Handedness Inventory [24], one subject was left handed. Foot dominance was assessed by asking subjects which foot they would preferably kick a ball with; all subjects were right foot dominant. The exclusion criteria were a history of back or hip problems or participants being less than 18 years of age.
2.2. Electromyography (EMG) and accelerometry Bilateral EMG activity was recorded from the external (EO) and internal obliques (IO), rectus abdominis (RA), erector spinae (ES) at the L4 vertebral level and from the right rectus femoris (RRF). The participants’ skin was first prepared using alcohol wipes then pairs of disposable silver/silver chloride electrodes (self-adhesive, ARBO blue, 2 cm diameter, Henleys Medical Supplied Ltd., Welwyn Garden City, UK) placed over the muscles in line with the orientation of the muscle fibres [25] with approximately 2 cm separation between the two electrodes of the pair. Ground electrodes were placed over the left patella, left olecranon and left ulnar styloid process. The EMG signals were filtered (10 Hz–1 kHz), amplified (1000, ISO-DAM bioamplifiers, World Precision Instruments, Stevenage, UK) and sampled (1401 plus and Signal software, Cambridge Electronic Design Limited, Cambridge, UK) at 2 kHz. An in-house constructed accelerometer was attached to the participants’ right thigh to measure the onset of leg movement, and over the manubrium (2 axes, Charge Amplifier Type 2635, Bruel & Kjaer UK Ltd., Stevenage, UK) to measure the antero-posterior body sway. A weight was attached to the ankle of the right leg, to facilitate APAs (1.2 kg; 1.45–2.40% of the participants’ weight).
2.3. Protocol Participants were seated upright on an exercise ball (Gymnic Plus 55 BRQ, Ledraplastic, Italy, maximum diameter 50 cm, inflated to ensure hip and knee flexion of 908) with their arms by their sides, not holding on to the ball (see Fig. 1). If they usually wore corrective glasses, they were instructed to wear them and look at a specific mark on the wall, which was positioned at eye level. Subjects were then asked to verify that they were able to do this. Baseline levels of EMG activity were recorded for 20 s with the participants sitting quietly on the ball. Following this, subjects completed three tasks, the order of which was computer randomised for each subject: 1. Quiet hip flexion (QF). 2. Hip flexion while counting up from 1 out loud (S1s). 3. Hip flexion while completing serial sevens down out loud from 997 (S7s). S7s is a cognitive task which involves the repeated subtraction of 7 from a given number, in this protocol 997. Subjects performed the S1s task in order to control for the effects of counting out loud during the S7s task; this would require a very low level of cognitive processing but the effects of speech would be controlled. For each task the subject was instructed to flex their hip as quickly as possible when they heard a pure tone (500 Hz, 0.1 s duration) delivered through a loud speaker at random time intervals. The tone was initiated by the experimenter pressing a button held behind and out of sight of the subject, to prevent anticipation. Sampling of EMG and accelerometer data was centred on the time of the tone presentation (0 ms in Fig. 2A), with a prestimulus period of recording of 250 ms (not shown). The experimenter pressed the button at the end of respiration as the phase of respiration is known to effect trunk muscle activation [26]. This was determined visually and achieved by the experimenter being positioned to the side of the subject. The tone volume was constant, but verified by each subject prior to starting the protocol, as being high enough that they could hear it but not too high so as to startle them. Within each task the subjects repeated the hip flexion ten times.
Please cite this article in press as: Jones P, et al. Effect of dual tasking on postural responses to rapid lower limb movement while seated on an exercise ball. Gait Posture (2014), http://dx.doi.org/10.1016/j.gaitpost.2014.04.185
G Model
GAIPOS-4193; No. of Pages 8 P. Jones et al. / Gait & Posture xxx (2014) xxx–xxx
3
Fig. 1. Experimental setup with circles denoting EMG electrode placement (anterior ones shown only) and squares denoting accelerometer placements.
2.4. Data processing and statistical analysis Data were analysed offline using Signal 3.13 (Cambridge Electronic Design Limited, Cambridge, UK). Using an automated script, all EMG data were first rectified and the mean (and two standard deviations [2SD]) of 100 ms duration of pre auditory cue level was calculated for each frame and denoted by 2 horizontal cursors (see Fig. 2A). Each frame was visually inspected to obtain the time at which RRF EMG activity exceeded the 2SD cursor, this was designated at time zero (see Fig. 2B). The automated script then measured the mean rectified EMG over 20 ms epochs from 250 ms before time zero to 350 ms after time zero for all muscles for each frame. The EMG activity during each 20 ms epoch was averaged over the 10 frames for each task and normalised to the activity obtained during the initial period of quiet sitting; a version of this method has been used previously [27]. Four time periods were established based on current literature [1]: APA1 ( 250 ms to 100 ms), APA2 ( 100 ms to +50 ms), CPA1 (+50 ms to +200 ms) and CPA2 (+200 ms to +350 ms) to facilitate identification of when the trunk muscles were becoming active. The data obtained from the two accelerometers was averaged using the computer automated script over the same time window and normalised to the activity obtained during the initial period of quiet sitting. Two way repeated measures ANOVA with factors time ( 250 ms to 350 ms in 20 ms epochs) and task (QF, S1s and S7s) was used to
compare the rectified EMG activity and the accelerometry data (Sigmplot 11.2, Systat Software, Chicago, USA). Within and between group comparisons were also made with appropriate correction for multiple comparisons, to establish differences within tasks across time and between tasks at each time point.
3. Results Tables 1 and 2 summarise the results of the ANOVA for the EMG and accelerometry data, respectively. The results of the current study show that the ES muscle contralateral to RRF showed a difference during the APA time window. There was a tendency for the onset of rise in EMG activity to be earlier in S7s than in S1s and QF. In addition, the EMG activity was higher in the S7s. This is in contrast to the effects in RF. The ES muscles ipsilateral to RF were more active during S7s and S1s than QF during the CPA time windows. In addition, the EO muscle contralateral to RRF was less active in S7s than either S1s or QF in the CPA time windows whereas ipsilateral to RRF were more active during S7s than QF during the CPA time windows. The RA and IO muscles showed no differences between the tasks in any APA or CPA windows. The rise in EMG activity in RRF became significantly different from baseline earliest in QF (20 ms after the auditory cue) and later in S1s and S7s (40 ms). Overall, EMG activity was significantly different between QF and S7s (P = 0.019) but not between QF and the S1s (P = 0.106). The overall mean level of EMG activity was highest for QF (11 times normalised baseline EMG activity), lower for S1s (9) and lowest for S7s (8.5) – see Fig. 2B. The epochs where there were significant differences between QF and S7s were 40–120 ms and 160–240 ms and between QF and S1s at 60, 80, 120 and 160 ms (p =
GAIPOS-4193; No. of Pages 8 Gait & Posture xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Gait & Posture journal homepage: www.elsevier.com/locate/gaitpost
Effect of dual tasking on postural responses to rapid lower limb movement while seated on an exercise ball P. Jones a, I. Sorinola b, P.H. Strutton a,* a The Nick Davey Laboratory, Human Performance Group, Division of Surgery, Department of Surgery and Cancer, Faculty of Medicine, Imperial College London, Charing Cross Hospital, London W6 8RF, UK b Division of Health and Social Care Research, School of Medicine, King’s College London, SE1 1UL, UK
A R T I C L E I N F O
A B S T R A C T
Article history: Received 13 March 2013 Received in revised form 3 April 2014 Accepted 8 April 2014
Postural adjustments are used by the central nervous system to pre-empt and correct perturbations in balance during voluntary body movements. Alteration in these responses is associated with a number of neuromuscular/musculoskeletal conditions. Attention has been identified as important in this system; performing a concurrent cognitive task has been suggested to reduce the efficacy of this postural control. The aim of this study was to examine the effect of concurrent cognitive tasking on anticipatory postural adjustments while sitting on an exercise ball with a view to help inform future rehabilitation programmes. Bilateral EMG activity was recorded from the external and internal obliques, rectus abdominis, erector spinae and the right rectus femoris of 20 healthy subjects (9 males) with mean (SD) age of 21.88 (0.86) years (range 21–24 years). A rapid hip flexion protocol was carried out under three conditions: no concurrent task, counting out loud up from one and completing a serial sevens task. The addition of the cognitive task delayed and reduced the EMG in the prime mover muscle but had little impact on the responses of the trunk muscles within the time frame of the anticipatory responses; suggestive of a decoupling of voluntary and postural control mechanisms. The results of this study suggest that perhaps the clinical effects of dual task may not be largely due to changes in anticipatory postural adjustments. However, it would be important to compare these results to those seen in older and functionally impaired individuals as this would be more representative of the typical population undertaking such rehabilitation programmes. ß 2014 Elsevier B.V. All rights reserved.
Keywords: Dual tasking Postural adjustments Trunk muscles Exercise ball Electromyography
1. Introduction Voluntary body movements generally displace the body’s centre of mass (COM) causing a disruption of posture [1,2]. In order to correct this shift in the COM, the central nervous system makes two overlapping phases of adjustments to trunk muscle activation. The first phase, termed anticipatory postural adjustments (APAs), begins before or is concurrent with the movement taking place; and aimed at minimising the postural effects of the movement [3]. These APAs are pre-planned and activated in a specific pattern by the CNS based on previous experience of the
* Corresponding author at: The Nick Davey Laboratory, Human Performance Group, Division of Surgery, Department of Surgery and Cancer, Faculty of Medicine, Imperial College London, Charing Cross Hospital, London W6 8RF, UK. Tel.: +44 (0) 203 313 8837; fax: +44 (0) 203 313 8835. E-mail address: [email protected] (P.H. Strutton).
effects of similar movements [4]. The second set of adjustments, compensatory postural adjustments (CPAs), counteract the actual perturbation in balance caused by the movement and are initiated by somatosensory feedback [5]. APAs are activated in the contexts of predictable perturbations such as self-generated movements; if the perturbations are unpredictable, APAs are often not observed [1]. CPAs, on the other hand, are required for both predictable and unpredictable perturbations and are initiated by somatosensory feedback. The mechanism of interaction between CPAs and APAs is not completely understood. However, there is evidence to suggest that stronger APAs reduce the need for large CPAs to preserve body equilibrium [1]. Deficits in postural adjustments [6] associated with limb movements, such as delayed activation and reduced magnitude of postural muscle activities have been described in a number of neuromuscular and musculoskeletal conditions [7,8], with consequences for functional abilities [9]. Therefore, strategies to improve APAs activations are a common construct in the
http://dx.doi.org/10.1016/j.gaitpost.2014.04.185 0966-6362/ß 2014 Elsevier B.V. All rights reserved.
Please cite this article in press as: Jones P, et al. Effect of dual tasking on postural responses to rapid lower limb movement while seated on an exercise ball. Gait Posture (2014), http://dx.doi.org/10.1016/j.gaitpost.2014.04.185
G Model
GAIPOS-4193; No. of Pages 8 2
P. Jones et al. / Gait & Posture xxx (2014) xxx–xxx
exercise programmes provided by physiotherapists. These programmes are often targeted at improving trunk activity and these can involve the use of exercise balls [10,11]. Although exercise balls are used extensively in motor rehabilitation the underlying mechanisms of any effects are still relatively unexplored. We have recently shown greater amplitudes of APAs and CPAs in trunk muscles occurring during lower limb movement while seated on an exercise ball compared to the same movement performed on a stable chair [12]. It was also suggested that there was a trend for early APA muscle activation while on the exercise ball compared to the chair [12]. These observations suggest that exercise balls might exert their influence on postural control system by modulation of the amplitudes and timing of APAs and CPAs. Postural control is often challenged during dual-tasking situations, when limb movements involved in activities of daily living tasks such as standing, walking and transfers are performed simultaneously with another task (i.e. cognitive or motor) [13]. To improve the ability to withstand these challenges, training under dual task conditions is often added to balance rehabilitation to facilitate better postural performance in these situations. Although improvements in postural control under dual task conditions [14] have been demonstrated with training, its effect on APAs and CPAs is still unclear. It is generally accepted that concurrent tasks and postural control either share available information processing capacity or the CNS organises their execution serially [15]. Therefore, if attentional resources are exceeded under dual tasking conditions, the performance of one or both tasks may become degraded. Previous studies have reported varying impacts of dual tasking on postural control. For example, increased stance sway was reported with concurrent performance of tasks with visuo-spatial components [16,17], or those that require verbal responses [18,19] or manipulation of surface and visual surround as in the Sensory Organisation Test [20,21]. In addition, it has also been found that the response EMG amplitude of postural muscles to perturbation was reduced under a dual task paradigm [22]. In contrast to these, other studies have found enhanced stance postural control [13,23] especially under dual task conditions with low cognitive demands. These observations suggest a varying impact on dual performance of cognitive tasks depending on the demands of both elements of the dual task paradigms. The effect dual tasking has on APAs and CPAs is currently unknown especially when combined with an exercise ball. Consequently, we sought to investigate the effect of dual tasking on postural adjustments to lower limb movements performed when seated on an exercise ball. We hypothesised that performing a simultaneous cognitive task will alter the magnitude and delay the timing of APAs and CPAs in trunk muscles while seated on an exercise ball.
2. Methods 2.1. Subjects With local ethical approval and written informed consent 20 healthy participants were recruited (9 males), with mean (SD) age of 21.88 (0.86) years (range 21–24 years), height 175.71 (9.15) cm and weight 69.71 (10.68) kg. Hand dominance was assessed using the Edinburgh Handedness Inventory [24], one subject was left handed. Foot dominance was assessed by asking subjects which foot they would preferably kick a ball with; all subjects were right foot dominant. The exclusion criteria were a history of back or hip problems or participants being less than 18 years of age.
2.2. Electromyography (EMG) and accelerometry Bilateral EMG activity was recorded from the external (EO) and internal obliques (IO), rectus abdominis (RA), erector spinae (ES) at the L4 vertebral level and from the right rectus femoris (RRF). The participants’ skin was first prepared using alcohol wipes then pairs of disposable silver/silver chloride electrodes (self-adhesive, ARBO blue, 2 cm diameter, Henleys Medical Supplied Ltd., Welwyn Garden City, UK) placed over the muscles in line with the orientation of the muscle fibres [25] with approximately 2 cm separation between the two electrodes of the pair. Ground electrodes were placed over the left patella, left olecranon and left ulnar styloid process. The EMG signals were filtered (10 Hz–1 kHz), amplified (1000, ISO-DAM bioamplifiers, World Precision Instruments, Stevenage, UK) and sampled (1401 plus and Signal software, Cambridge Electronic Design Limited, Cambridge, UK) at 2 kHz. An in-house constructed accelerometer was attached to the participants’ right thigh to measure the onset of leg movement, and over the manubrium (2 axes, Charge Amplifier Type 2635, Bruel & Kjaer UK Ltd., Stevenage, UK) to measure the antero-posterior body sway. A weight was attached to the ankle of the right leg, to facilitate APAs (1.2 kg; 1.45–2.40% of the participants’ weight).
2.3. Protocol Participants were seated upright on an exercise ball (Gymnic Plus 55 BRQ, Ledraplastic, Italy, maximum diameter 50 cm, inflated to ensure hip and knee flexion of 908) with their arms by their sides, not holding on to the ball (see Fig. 1). If they usually wore corrective glasses, they were instructed to wear them and look at a specific mark on the wall, which was positioned at eye level. Subjects were then asked to verify that they were able to do this. Baseline levels of EMG activity were recorded for 20 s with the participants sitting quietly on the ball. Following this, subjects completed three tasks, the order of which was computer randomised for each subject: 1. Quiet hip flexion (QF). 2. Hip flexion while counting up from 1 out loud (S1s). 3. Hip flexion while completing serial sevens down out loud from 997 (S7s). S7s is a cognitive task which involves the repeated subtraction of 7 from a given number, in this protocol 997. Subjects performed the S1s task in order to control for the effects of counting out loud during the S7s task; this would require a very low level of cognitive processing but the effects of speech would be controlled. For each task the subject was instructed to flex their hip as quickly as possible when they heard a pure tone (500 Hz, 0.1 s duration) delivered through a loud speaker at random time intervals. The tone was initiated by the experimenter pressing a button held behind and out of sight of the subject, to prevent anticipation. Sampling of EMG and accelerometer data was centred on the time of the tone presentation (0 ms in Fig. 2A), with a prestimulus period of recording of 250 ms (not shown). The experimenter pressed the button at the end of respiration as the phase of respiration is known to effect trunk muscle activation [26]. This was determined visually and achieved by the experimenter being positioned to the side of the subject. The tone volume was constant, but verified by each subject prior to starting the protocol, as being high enough that they could hear it but not too high so as to startle them. Within each task the subjects repeated the hip flexion ten times.
Please cite this article in press as: Jones P, et al. Effect of dual tasking on postural responses to rapid lower limb movement while seated on an exercise ball. Gait Posture (2014), http://dx.doi.org/10.1016/j.gaitpost.2014.04.185
G Model
GAIPOS-4193; No. of Pages 8 P. Jones et al. / Gait & Posture xxx (2014) xxx–xxx
3
Fig. 1. Experimental setup with circles denoting EMG electrode placement (anterior ones shown only) and squares denoting accelerometer placements.
2.4. Data processing and statistical analysis Data were analysed offline using Signal 3.13 (Cambridge Electronic Design Limited, Cambridge, UK). Using an automated script, all EMG data were first rectified and the mean (and two standard deviations [2SD]) of 100 ms duration of pre auditory cue level was calculated for each frame and denoted by 2 horizontal cursors (see Fig. 2A). Each frame was visually inspected to obtain the time at which RRF EMG activity exceeded the 2SD cursor, this was designated at time zero (see Fig. 2B). The automated script then measured the mean rectified EMG over 20 ms epochs from 250 ms before time zero to 350 ms after time zero for all muscles for each frame. The EMG activity during each 20 ms epoch was averaged over the 10 frames for each task and normalised to the activity obtained during the initial period of quiet sitting; a version of this method has been used previously [27]. Four time periods were established based on current literature [1]: APA1 ( 250 ms to 100 ms), APA2 ( 100 ms to +50 ms), CPA1 (+50 ms to +200 ms) and CPA2 (+200 ms to +350 ms) to facilitate identification of when the trunk muscles were becoming active. The data obtained from the two accelerometers was averaged using the computer automated script over the same time window and normalised to the activity obtained during the initial period of quiet sitting. Two way repeated measures ANOVA with factors time ( 250 ms to 350 ms in 20 ms epochs) and task (QF, S1s and S7s) was used to
compare the rectified EMG activity and the accelerometry data (Sigmplot 11.2, Systat Software, Chicago, USA). Within and between group comparisons were also made with appropriate correction for multiple comparisons, to establish differences within tasks across time and between tasks at each time point.
3. Results Tables 1 and 2 summarise the results of the ANOVA for the EMG and accelerometry data, respectively. The results of the current study show that the ES muscle contralateral to RRF showed a difference during the APA time window. There was a tendency for the onset of rise in EMG activity to be earlier in S7s than in S1s and QF. In addition, the EMG activity was higher in the S7s. This is in contrast to the effects in RF. The ES muscles ipsilateral to RF were more active during S7s and S1s than QF during the CPA time windows. In addition, the EO muscle contralateral to RRF was less active in S7s than either S1s or QF in the CPA time windows whereas ipsilateral to RRF were more active during S7s than QF during the CPA time windows. The RA and IO muscles showed no differences between the tasks in any APA or CPA windows. The rise in EMG activity in RRF became significantly different from baseline earliest in QF (20 ms after the auditory cue) and later in S1s and S7s (40 ms). Overall, EMG activity was significantly different between QF and S7s (P = 0.019) but not between QF and the S1s (P = 0.106). The overall mean level of EMG activity was highest for QF (11 times normalised baseline EMG activity), lower for S1s (9) and lowest for S7s (8.5) – see Fig. 2B. The epochs where there were significant differences between QF and S7s were 40–120 ms and 160–240 ms and between QF and S1s at 60, 80, 120 and 160 ms (p =
