External and internal focus of attention differentially modulate corticospinal excitability in anticipatory postural adjustments

In the present study, we investigated the effects of attentional focus on corticospinal excitability of the lower limb muscles in the APA phase using a complex and goal-directed motor task of dart throwing. The major finding was that the corticospinal excitability of the TA muscle significantly increased prior to muscle contraction in both EF and IF conditions, and that the increment was significantly greater in the EF condition than in the IF condition. To the best of our knowledge, the present study provides the first evidence that attentional focus may play an important role in the modulation of corticospinal excitability in APA.

Motor performance during APA with different attentional foci

APA with lower limb, upper limb, and trunk muscles operates for human voluntary movements, and most previous studies used simple and/or single-joint movements to investigate APA18,19,20,21. We adopted a more complex and goal-directed movement of dart throwing to ensure that participants easily changed their attentional focus with the movement. The resulting performance accuracy was significantly greater in the EF condition than in the IF condition, in line with previous studies1. Subsequently, it is of interest to determine whether joint movements of the upper and lower limbs and COP during APA differ depending on the focus of attention applied.

The COP shifted in a posterior direction at a significantly early stage accompanied by a burst of TA muscle after the visual cue (Table 1), and it might be triggered in advance by slight elbow flexion before full elbow extension, as previously reported29. The angular velocity of elbow extension was greater in the IF condition than in the EF condition, and this is reasonable in terms of attentional focus. Because participants concentrated on the elbow joint in the IF condition, they might throw a dart by mainly extending the elbow joint, resulting in acceleration of the angular velocity of the elbow extension with less performance accuracy (Fig. 2), according to the theory of speed-accuracy trade-off33. An increase in the velocity of the intended movement might affect APA in duration and/or magnitude, as suggested by a previous study34. Conversely, the changes in the angular velocity of the hip and knee joints and the change in the angle of the ankle joint were slightly greater in the EF condition than in the IF condition. These results suggest that participants throw the dart using whole-body movement with emphasis on the lower limb when they concentrate on the flight trajectory of the dart in the EF condition. Interestingly, despite the differences in the kinematic results of the upper and lower limbs, the COP showed no significant differences between the EF and IF conditions. Because the displacement of COP reflects whole-body movement, including the upper limb, trunk, and lower limb, it is likely that the perturbations from the throwing movement and postural control were counteracted and showed no difference between the EF and IF conditions.

EMG activities during APA with different attentional foci

Previous studies have reported that the agonist and/or antagonist muscles showed less activity under EF than IF conditions in force production35,36 and dart throwing task37. In the present study, the AD muscle (synergist) showed a trend of less activity in the EF condition than in the IF condition, suggesting that EF might lead to efficient coordination of the agonist and/or synergist muscles and might result in better performance accuracy and less muscular activity of the AD muscle in dart throwing.

The upper limb movement of dart throwing including a slight elbow flexion followed by a full elbow extension would cause a slight posterior followed by an anterior shift of COM. Prior to these upper limb movements, COP would shift in a posterior and an anterior direction, respectively. As shown in the Fig. 4 and Table 1, the TA muscle activity and subsequent posterior COP shift started preceding the elbow flexion, which caused the anterior shift of COM. This preceding anterior-shifted COM might be responsible for counteracting or minimising the upcoming posterior shift of COM by the elbow flexion to maintain the postural stability. At this early stage, the TA muscle activity might be an anticipatory postural reaction to stabilize the body sway from the elbow flexion of dart throwing in advance. Thereafter, the deactivation of the TA muscle and/or activation of the SOL muscle caused the anterior-shift of COP and decelerated the anterior-shift of COM, namely a posterior-shift of COM. These sequential changes in the TA and SOL muscle activities as well as COP would operate for counteracting the upcoming, strong and anterior shift of COM by elbow extension. In particular, the SOL muscle onset was observed after the elbow flexion movement and showed no significant difference from the TB muscle onset, suggesting that the activity of this antigravity muscle was interpreted as a compensatory, but not an anticipatory, postural reaction during dart throwing.

In spite of these sequential and complex COP-COM dynamics during dart throwing, there was no significant difference in burst timing and amount of TA or SOL muscle between the EF and IF conditions. At a performance level, previous work utilising a Fitts’ task in the lower limb (i.e. a fast one-leg movement) showed that EF led to better motor performance, longer APA duration, and smaller APA magnitude than IF38,39, suggesting that the standing posture might be controlled more efficiently when adopting an EF strategy. The difference in motor task or interlimb/intralimb coordination might explain the difference between the previous and present results. In a previous study, the postural muscles were directly involved in the motor task, whereas a complex and goal-directed upper limb movement was used to explore the APA in the lower limb muscles, which were not directly involved in the motor task in the present study. There were no significant differences in the APA duration and magnitude between the EF and IF conditions, whereas the kinematic results showed greater changes in the angle and angular velocity of the lower limb in the EF condition, as mentioned above. One explanation for these observations is that the comparable magnitude of lower limb EMG activity with larger joint movements with an EF strategy might lead to more effective postural control than an IF strategy during dart throwing. In addition to the TA and SOL muscles, other lower limb muscles for extension of the hip joint and/or flexion of the knee joint might contribute to the significance between the EF and IF conditions. This is purely hypothetical, and further studies are required to address this issue.

Corticospinal excitability in APA with different attentional foci

Consistent with our previous report29, the corticospinal excitability of the TA muscle, which contributes to APA in dart throwing, significantly increased before the EMG burst under both EF and IF conditions. The corticospinal excitability of the SOL muscle showed no change, suggesting that the corticospinal tract play a role in APA in a muscle-dependent manner. Importantly, MEP increments in the TA muscle prior to TA onset and at TB onset were significantly larger in the EF condition than in the IF condition (Fig. 5), indicating differential regulation of corticospinal excitability immediately prior to and during the APA phase. Considering that the angle and angular velocity changes in the lower limb were significantly greater in the EF condition than in the IF condition, the increased corticospinal excitability of the TA muscle in the EF condition would contribute to controlling the subsequent greater movement of the lower limb during throwing movement.

Previous studies using TMS reported no significant difference in the corticospinal excitability of the agonist muscle between the EF and IF conditions in finger16,17,40 and elbow41 movements. However, the short-interval intracortical inhibition in the M1 of the agonist muscle was greater in the EF condition than in the IF condition, and the levels of surround inhibition in the adjacent muscle were higher when adopting an EF compared to an IF during force control16,17. Moreover, adopting an IF strategy showed higher excitability of slow motor pathways, but not fast motor pathways, in comparison with the EF strategy40. Recently, an electroencephalogram (EEG) study further revealed that the IF condition increased the EEG coherence (10–12 Hz) of Alpha 2 between T3 (verbal-analytical region) and Fz (motor planning region) compared to that without any instruction regarding attentional focus, suggesting that an IF strategy might lead to higher real-time conscious motor processing and reduce the accuracy of motor performance42. These results highlighted a dissociation between corticospinal and cortical excitability corresponding to the agonist muscle and suggested that better motor performance in the EF condition was accompanied by enhanced levels of cortical inhibition and a more efficient neural strategy. Our results further revealed that the corticospinal excitability of the postural muscles engaged in APA increased, suggesting that attentional focus potentially modulates corticospinal excitability in the APA muscle. The larger increment of MEP in the EF condition compared with the IF condition may reflect a stronger neural drive (central motor command) originating in higher brain centres.

Multiple descending pathways link the cortex to the spinal cord, enabling the transmission of central motor commands for voluntary movement to spinal motoneurons. MEP to TMS mainly reflects the excitability of the corticospinal tract at cortical and spinal levels. Using single-pulse TMS without direct measurement of cortical or spinal excitability separately, it is difficult to identify the origin responsible for our results. Previous studies using combinations of single- and paired-pulse TMS, cervicomedullary stimulation, and H-reflex have shown that the M1 would be involved in APA control21,26. In addition, the MEP of the TA muscle in the present study were clearly facilitated in the absence of EMG activity prior to TA onset, and a distinct difference was observed between the EF and IF conditions. The amount of MEP enhancement prior to TA onset without EMG activity was comparable to that at TB onset, in which EMG activities were always involved, suggesting an excitability change at the supraspinal level rather than at the spinal level. Taking into consideration the previous and present results, it is most likely that the APA in the lower limb was pre-programmed by the CNS, and the differential enhancements of MEP depending on the cognitive state (EF and IF conditions) might be attributed to excitability changes at the supraspinal level, such as the M1. Other cortical regions involving the primary somatosensory cortex, premotor cortex, supplementary motor area, insular cortex, and intraparietal lobule are also candidates for contributing to the difference between the EF and IF conditions12,13,14. Interestingly, the supplementary motor area is also involved in APA43,44, and it may play a role in mediating the M1 or corticospinal excitability during APA with different attentional foci. Therefore, future neuroimaging studies are warranted.

Apart from the corticospinal tract or cerebral cortex, the reticulospinal tract is a candidate for the modulation of motor neuron excitability. In the literature, the primate reticulospinal tract is usually considered to control proximal and axial muscles and is involved mainly in gross movements, such as locomotion, reaching, and posture45. The reticular formation of the brain stem processing sensory input and guiding motor output is hypothesised to be responsible for APA. During voluntary movements, because the central motor command activates the spinal motor neuron and the neural circuits in the brain stem concomitantly, the increased M1 excitability by the voluntary drive of dart throwing may activate the spinal motoneuron pool of the TA muscle directly through the corticospinal tract and indirectly through the corticoreticular tract.


The present study has some limitations that should be acknowledged. First, no control condition (without instruction of attentional focus) was adopted in the present study. Our recent study29 has suggested that it was difficult to make a neutral focus condition, because the participants paid attention with the motor task either on the movement outcome and external environment (EF) or their own body and movement (IF) consciously or unconsciously. To elucidate whether the corticospinal excitability is modulated differentially by EF and IF, we adopted the EF and IF conditions in order to distinguish the attentional focus condition clearly in the present study. Second, the motor performance involving kinematic and EMG profiles during dart throwing was measured in a separate protocol without TMS to precisely determine the behavioural difference between the EF and IF conditions. TMS using the double cone coil would affect the following motor performance because the stimulation would spread in the M1 and induce muscle activation in other trunk and/or upper limb muscles apart from the TA and SOL muscles. Third, it was difficult to precisely clarify the temporal changes in corticospinal excitability in the SOL muscle because TMS was applied over the motor hotspot of the TA muscle and stimulus timings were determined by the TA muscle onset. According to our results, the corticospinal excitability of the SOL muscle remained unchanged immediately before TA muscle onset, whereas that of the TA muscle increased. Finally, because the cortical and spinal excitability were not measured separately, it was difficult to identify the origin responsible for our results. According to our results and previous studies16,17,21,26, it is likely that differential modulations of the corticospinal excitability between the EF and IF conditions in APA are originated from the excitability changes at the cortical level. Further studies are required to clarify this.