The influence of normal human ageing on automatic movements

Abstract

There is evidence that aged normal subjects have more difficulty in achieving automaticity than young subjects. The underlying central neural mechanism for this phenomenon is unclear. In the present study, functional magnetic resonance imaging (fMRI) was used to investigate the effect of normal ageing on automaticity. Aged healthy subjects were asked to practice self-initiated, self-paced, memorized sequential finger movements with different complexity until they could perform the tasks automatically. Automaticity was evaluated by having subjects perform a secondary task simultaneously with the sequential movements. Although it took more time, most aged subjects eventually performed the tasks automatically at the same level as the young subjects. Functional MRI results showed that, for both groups, sequential movements activated similar brain regions before and after automaticity was achieved. No additional activity was observed in the automatic condition. While performing automatic movements, aged subjects had greater activity in the bilateral anterior lobe of cerebellum, premotor area, parietal cortex, left prefrontal cortex, anterior cingulate, caudate nucleus and thalamus, and recruited more areas, including the pre-supplementary motor area and the bilateral posterior lobe of cerebellum, compared to young subjects. These results indicate that most healthy aged subjects can perform some complex motor tasks automatically. However, aged subjects appear to require more brain activity to perform automatically at the same level as young subjects. This appears to be the main reason why aged subjects have more difficulty in achieving automaticity.

Normal ageing is not only characterized by a decline of memory, perception and cognition (Mark & Rugg, 1998; Grady, 2000), but is also accompanied by progressive slowness and impaired motor ability (Sailer et al. 2000; Calautti et al. 2001; Mattay et al. 2002). Some functional imaging studies found more activation in aged subjects in several brain areas compared to young subjects while performing the same simple motor tasks (Calautti et al. 2001; Mattay et al. 2002; Hutchinson et al. 2002; Ward & Frackowiak, 2003). In contrast, a study on motor sequence learning found no ageing-related reorganization (Daselaar et al. 2003). These results suggest that normal ageing may have different effects in various motor tasks. However, most previous functional imaging experiments employed simple tasks, and the influence of ageing on more complex motor behaviours has been only rarely studied.

A general characteristic of the motor system is that people can perform some learned movements automatically. Such movements are performed without attention being clearly directed towards the details of the movement (Bernstein, 1967). For example, musicians can perform music accurately while holding a conversation. For young subjects, after a period of training even some complex tasks can be executed automatically (Wu et al. 2004). It has been speculated that aged adults have more difficulty achieving automaticity than younger subjects (Rogers et al. 1994). However, it is unclear whether healthy aged subjects can achieve full automaticity after sufficient practice. Moreover, if aged subjects can achieve automaticity, it is not clear whether there would be a difference in the pattern of brain activation.

In the present study, we used several sequential movement tasks with different complexity to investigate the influence of normal ageing on automaticity. To evaluate whether automaticity was achieved, we applied a dual task protocol (Wu et al. 2004). The evidence that a task has become automatic can be proven by the fact that a secondary task can be performed simultaneously with minimal or no interference (Passingham, 1996). The first part of the dual task required subjects to perform some self-initiated, self-paced, memorized sequential finger tapping movements. The secondary part was a letter-counting task in which subjects were asked to identify the number of times a target letter from a sequence of letters was seen.

We employed functional magnetic resonance imaging (fMRI) to observe the related brain activity. A recent neuroimaging study on normal young subjects by our group revealed that most of the motor network participates in executing automatic movements and some areas, such as the bilateral cerebellum, anterior supplementary motor area (SMA), cingulate cortex, left caudate nucleus, bilateral premotor cortex, bilateral parietal cortex and dorsal lateral prefrontal cortex become less activated as movements become more automatic. In addition, we did not find any area that was more activated for automatic movements (Wu et al. 2004).

The objective of this study was to compare performance and brain activation related to the process of automaticity in aged and young normal individuals. We hypothesized that the aged subjects require more brain network recruitment than the young subjects to execute automatic movements. We also made additional observations on the influence of normal ageing on performance of dual tasks.

Methods

Subjects

We studied 14 healthy aged subjects. Two subjects were excluded because they did not achieve automaticity after training. The remaining 12 subjects ranged in age from 57 to 73 years old (mean 61.8 years), and included eight males and four females. We also investigated 12 sex-matched young subjects, aged 23–38 years (mean 30.5 years). The results from these young subjects were previously reported (Wu et al. 2004). The subjects were all right-handed (laterality index was 0.8–1.0) as measured by the Edinburgh Inventory. The aged group was also administered the Mini-Mental State Exam (MMSE). The MMSE score was 30 in all subjects. They reported no history of neurological illness or psychiatric history. No subjects were taking any medications that could affect brain excitability. For all subjects, no significant pathological change was found with standard T1- and T2-weighted MRI, although there were changes due to normal ageing. The experiments were performed according to the Declaration of Helsinki and were approved by the Institutional Review Board. All subjects gave their written informed consent for the study.

Methods

All procedures were identical to those of our previous paper (Wu et al. 2004) and are only briefly described here. Subjects were asked to perform two sequences of right finger tapping, referred to as sequence-4 and sequence-12, based on the number of movements in each unit of the sequence. ‘Sequence-4’ was 1–3–4–2, in which 1, 2, 3 and 4 refer to the index, middle, ring and little fingers, respectively. ‘Sequence-12’ was 1–4–3–2–2–4–1–3–4–1–2–3. Both sequences were executed at 0.5 Hz. Automaticity was evaluated by having subjects perform a visual letter-counting task simultaneously with these sequential movements. For the letter-counting task, letter sequences consisting of a random series of the letters A, G, L and O were presented on a screen and subjects were asked to identify the number of times they saw a specified target letter. All sequential movements were self initiated and self paced. Before the first scan, all subjects practised until they could move at the required rate. They briefly practiced each sequential movement. In addition, subjects were given enough practice trials to ensure that they could perform the visual letter-counting tasks correctly with no difficulty. After scanning, subjects were asked to report the number of target letters. After the first scan, subjects practised these tasks until they could perform sequence-4 and sequence-12 from memory 10 times in a row without error, as well as the dual tasks accurately.

Control experiment

In order to explain whether the age-related change in brain activity was due to difference of strength, a control study was carried out. All aged and young subjects were asked to tap their left index finger at a frequency of 1 Hz. This tapping task was a self-paced task and was well trained before the fMRI scan.

Functional MRI experiments

T2* time constant-sensitive functional images were obtained using a whole-body 1.5 T magnetic resonance imaging (MRI) scanner (Signa, General Electric, Milwaukee, WI, USA) and a standard head coil. Subjects lay supine in the MR scanner with a response device fixed to their hands. The response device had five buttons, corresponding to the index, middle, ring and little fingers of the right hand; and the index finger of the left hand and was used to record finger movements. The subjects viewed visual signals on a screen through a mirror built into the head coil. We used an echo planar imaging (EPI) gradient echo sequence (21 slices, echo time (TE) = 30 ms, repetition time (TR) = 2500 ms, flip angle = 90 deg, field of view (FOV) = 22 × 22 cm, matrix = 64 × 64) to obtain functional images. A time-course series of 100 images per slice was acquired for each trial, in an off–on cycle protocol of rest and activation. Each scanning session lasted 4 min.

Functional MRI scans were acquired both before and after the subjects achieved automaticity. Two conditions were contained in each scanning session and were defined as the ‘rest’ and ‘active’ conditions. Each condition lasted 25 s and was repeated five times in a session. In the rest condition, subjects were asked to relax and focus on the screen in front of them. The active condition in each session contained one of the six single or dual tasks: sequence-4, sequence-12, letter counting, left hand tapping, dual task of sequence-4 and letter counting, and dual task of sequence-12 and letter counting. No feedback was provided during scanning to tell subjects whether their finger movements were correct or incorrect.

Behavioural data analysis

Each subject's performance for each task was recorded. Errors were used to evaluate whether these tasks were performed automatically after training. Only the performances achieving high accuracy in both single and dual tasks were considered automatic. The performance of each task of the aged group was compared to the young subjects (two-sample t test, P < 0.05). Within each group, the difference in performance before and after automaticity was achieved, and between single and dual task, was calculated (repeated-measures ANOVA, P < 0.05). In addition, the performance between sequence-12 and sequence-4 was compared (two-sample t test, P < 0.05).

Imaging data analysis

Image analysis was performed with SPM 99 software (Wellcome Institute of Cognitive Neurology, London, UK). Functional images were aligned to the first image of each session for motion correction. After spatial normalization, all images were resampled into voxels that were 2 × 2 × 2 mm in size. Images were also smoothed with a Gaussian filter of 6 mm full-width at half-maximum (FWHM). Both first- and second-level analyses were performed. In the first-level analysis, data were analysed for each single subject separately on a voxel-by-voxel basis using the principles of the general linear model extended to allow the analysis of fMRI data as a time series. The data were modelled using a fixed effect boxcar design, convolved with a haemodynamic response function chosen to represent the relationship between neuronal activation and blood flow changes. The model had the same on–off frequency as the alternation frequency of the active and rest conditions, and was constructed for analysis of task-dependent activation, identical for all subjects and for all conditions. A contrast representing the effect of the active condition compared with the rest condition was defined and contrast images were calculated individually for each condition. In addition, a contrast representing ‘deactivation’, which means more voxel intensity during rest condition than during active condition, was also calculated. These contrast images were used in the second-level analysis for random effects. For the within-group analysis, a one-sample t test model was used to identify the brain activity before and after training for each condition (P < 0.001, without correction for multiple comparisons). A student's paired t test model was used to compare the pretraining results with the post-training results for each condition (P < 0.001, uncorrected). In addition, one-way ANOVA was used to compare the results of dual tasks and single tasks (P < 0.001, uncorrected). For between-group comparisons, a two-sample t test model (P < 0.001, uncorrected) was used to explore the difference between aged and young subjects after training. We chose this threshold because it is often more informative and may show a trend towards increased activation, although not reaching the more conservative corrected statistical threshold. Locations of activated areas for different conditions were displayed by superimposing them on the Montreal Neurological Institute (MNI) template.

Results

Task performance

The accuracy of sequential finger movements and visual letter counting for single sequential movements and dual tasks across all aged and young subjects is shown in Table 1. Two aged subjects performed the dual task of sequence-4 and letter counting with high accuracy but could not perform sequence-12 and letter counting correctly after extensive training, which suggests that they could not achieve automaticity in performing sequence-12. Therefore, all their data were excluded. Before training, both groups committed errors with sequential finger movements in performing all sequential movements and dual tasks. The wrong finger taps of the sequences were consistent across the five blocks during the first scanning session. In both the aged and young groups, there were more finger movement errors in performing sequence-12 than in performing sequence-4 (two-sample t test, P < 0.05), and in performing dual tasks than in performing single tasks (ANOVA, P < 0.05). In addition, more errors were found when performing the dual task of sequence-12 and letter counting than when performing the dual task of sequence-4 and letter counting (ANOVA, P < 0.05). Aged subjects made significantly more errors than young subjects while performing dual tasks (ANOVA, P < 0.05). They also had more errors than young subjects in performing either sequence-4 or sequence-12, although the difference was not statistically significant (two-sample t test, P > 0.05).

Table 1.

Performance (percentage of errors) of sequential finger movements and dual task before and after training in aged and young subjects

	Aged subjects		Young subjects
Task	Errors before training (%)	Errors after training (%)	Errors before training (%)	Errors after training (%)
Sequence-4	4.8 ± 6.4	0	1.4 ± 2.6	0
Sequence-12	18.5 ± 13.2	1.1 ± 1.8	11.8 ± 15.5	0.6 ± 1.7
Letter counting	0	0	0	0
Sequence 4/letter counting	11.9 ± 12.8/9.9 ± 10.1	0.3 ± 0.8/1.2 ± 2.3	4.9 ± 4.2/5.9 ± 9.9	0/0.7 ± 1.9
Sequence-12/letter counting	30.5 ± 19.6/18.7 ± 11.9	1.2 ± 1.9/2.0 ± 2.6	17.1 ± 15.2/13.9 ± 11.5	1.1 ± 1.9/1.2 ± 2.2

Values are given as means ±s.d. for percentage of errors. The results of the dual task of sequential movements and visual letter counting are given as errors of finger movements/errors of letter counting.

Aged subjects needed significantly more time than young subjects to achieve automaticity (4.7 ± 1.0 versus 3.0 ± 0.7 h). After training, the performance of both aged and young groups was significantly improved (repeated-measures ANOVA, P < 0.05). They could perform single and dual tasks correctly. Moreover, both groups reported that they could execute the tasks without paying attention to the sequential finger movements and did not feel difficulty any more. There were no more errors in performing sequence-12 than in performing sequence-4, or in performing dual tasks than in performing single tasks in each group, and there was no significant difference between the two groups.

Additionally, there was no between- or within-group difference for the rate of sequential movements. Before and after training, the rates of aged subjects were 0.54 ± 0.07 and 0.52 ± 0.06 Hz, respectively, while the rates of young subjects were 0.55 ± 0.04 and 0.52 ± 0.03 Hz, respectively. However, during practice aged subjects reported more difficulty than young subjects in acquiring the required rate.

Within-group analysis of brain activity while performing single tasks

Before training, for aged subjects the performances of sequence-4 and sequence-12 were associated with activations in the left primary sensorimotor cortex, bilateral premotor areas, bilateral parietal cortex, bilateral inferior frontal gyrus, bilateral dorsal lateral prefrontal cortex, supplementary motor area (SMA)-proper, pre-SMA, anterior cingulate motor cortex, basal ganglia, bilateral insular cortex and bilateral cerebellum. After training, the pattern of brain activity was similar to that before training and no additional activation was observed for either sequence-4 or sequence-12 (Fig. 1). These results were similar to those for young subjects (Wu et al. 2004). In aged subjects, after training there was less activation in the bilateral premotor area, bilateral superior and inferior parietal lobes and pre-SMA compared to the before-training stage (P < 0.001, uncorrected; Fig. 2). There was less activity in bilateral cerebellum, bilateral premotor area, bilateral superior and inferior parietal lobes, left dorsal lateral prefrontal cortex, pre-SMA, anterior cingulate motor cortex and left caudate nucleus in the young group after training (Wu et al. 2004).

Figure 1. Brain areas activated during performance of automatic movements.

Brain regions activated during performing sequence-12 at the automatic stage in aged subjects. Results were thresholded at P < 0.001 (uncorrected).

Figure 2. Brain areas less activated in the automatic stage.

Brain areas more activated at the pretraining stage than at the automatic stage during the performance of sequence-12 in aged subjects. Results were thresholded at P < 0.001 (uncorrected) and rendered over a standard anatomical brain.

Between-group analysis of brain activity while performing single tasks

Areas more activated in the aged group at the after-training stage.

Compared to young subjects, at the after-training stage, aged subjects had greater activation in the bilateral cerebellum (anterior lobe), bilateral premotor area, bilateral parietal cortex, left dorsal lateral prefrontal cortex, anterior cingulate motor cortex, right caudate nucleus, thalamus and occipital cortex during performance of sequence-4 and sequence-12. Furthermore, the pre-SMA and bilateral posterior lobe of cerebellum, areas that were not activated in the young subjects any more at this stage, were still recruited in the aged subjects (Fig. 3 and Table 2).

Figure 3. Brain areas more activated in aged subjects.

Brain areas more activated in aged subjects than in young subjects during automatic execution of sequence-12 (P < 0.001, uncorrected).

Table 2.

Brain areas more activated in aged subjects than in young subjects while performing sequence-12 at the automatic stage

Cluster size	Activated areas	x	y	z	Z-value
398	R Caudate nucleus	2	2	9	6.28
1023	L Cerebellum, anterior lobe	−4	−56	−4	6.16
	L Cerebellum, posterior lobe	−20	−67	−27	5.80
597	L Cingulate	−8	29	4	6.12
156	R Parietal cortex	30	−53	54	6.08
339	L Pre-SMA	−2	14	51	6.06
529	L Premotor area	−39	2	48	5.93
29	R Premotor area	20	12	55	5.80
128	R Thalamus	14	−6	−1	5.78
71	L Parietal cortex	−50	−40	44	5.67
107	R Prefrontal cortex	42	25	−6	5.61
221	R Cerebellum, posterior lobe	37	−63	−20	5.57
	R Cerebellum, anterior lobe	32	−56	−27	5.38
29	L Precuneus	−6	−63	58	5.49
115	R Occipital cortex	6	−85	15	5.45
140	L Prefrontal cortex	−2	41	46	5.34

The coordinates are given as stereotaxic coordinates referring to the atlas of Talairach and Tournoux. Cluster size is the number of voxels. All areas were significant at P < 0.001 (uncorrected). Abbreviations: L, left; R, right; SMA, supplementary motor area; Z-value, Z-statistic value.

Areas less activated in the aged group

There was less activation in the left primary sensorimotor cortex (SM1) in the aged subjects than in the young subjects while performing sequence-4 and sequence-12 after training (Fig. 4). In contrast, in the control study (left index finger tapping), neither right nor left SM1 was less activated in the aged subjects compared to the young group.

Figure 4. Brain areas more activated in young subjects.

Brain regions more activated in young subjects than in aged subjects during performance of sequence-12 after training (P < 0.001, uncorrected).

Deactivations.

The comparison of brain deactivation during execution of sequence-4 and sequence-12 at the after-training stage between the two groups is shown in Fig. 5. Young subjects had more deactivation in the prefrontal, anterior cingulate and precuneus areas compared to aged subjects.

Figure 5. Brain areas more deactivated in young subjects.

Brain regions more deactivated in young subjects compared to aged subjects during automatic execution of sequence-12 (P < 0.001, uncorrected).

Within-group analysis of brain activity while performing dual tasks

Before training, for both groups, the performance of the dual task of sequence-4 and letter counting or sequence-12 and letter counting was associated with activations of the left primary sensorimotor cortex, bilateral premotor areas, bilateral parietal cortex, bilateral inferior frontal gyrus, bilateral dorsal lateral prefrontal cortex, SMA-proper, pre-SMA, cingulate cortex, basal ganglia, bilateral insular cortex, bilateral cerebellum and occipital cortex. Aged subjects also activated bilateral precuneus. After training, no additional activation was observed for either group. Similar to single tasks, after training there was less activation in the bilateral premotor area, bilateral parietal cortex and pre-SMA compared to the before-training stage in aged subjects. In young subjects, there was less activity in bilateral cerebellum, bilateral premotor area, bilateral parietal cortex, left dorsal lateral prefrontal cortex, pre-SMA, anterior cingulate motor cortex and left caudate nucleus. In aged subjects, additional activity was found in the bilateral precuneus during performance of dual tasks compared to the component tasks (one-way ANOVA, P < 0.001, uncorrected; Fig. 6). In young subjects, there was no brain area additionally activated for dual tasks.

Figure 6. Brain area specifically activated in dual tasks in aged subjects.

Brain area more activated in the dual task of sequence-12 and letter counting compared to single component tasks in aged subjects (one-way ANOVA, P < 0.001, uncorrected).

Between-group analysis of brain activity while performing dual tasks

At the after-training stage, aged subjects had greater activation in the bilateral cerebellum, bilateral premotor area, bilateral parietal cortex, bilateral precuneus, left dorsal lateral prefrontal cortex, pre-SMA, anterior cingulate motor cortex, caudate nucleus, thalamus and occipital cortex compared to young subjects during performance of dual tasks of sequence-4 and letter counting or sequence-12 and letter counting. In addition, we found less activation in the left SM1 in the aged subjects compared to the young subjects while performing dual tasks.

Discussion

Effect of ageing on automaticity

Before training, both groups made errors and the aged subjects made more errors than the young subjects. Two aged subjects failed to achieve automaticity, and the other aged subjects needed more training time than young subjects. These results demonstrate that aged subjects had greater difficulty than young subjects in achieving automaticity (Rogers et al. 1994). However, extensive practice was beneficial not only for young subjects, but also for aged subjects. Eventually most aged subjects significantly improved their performance and could execute these complex sequential movements automatically at the same level as the young subjects.

Aged subjects had a similar pattern of cortical activity for both the before- and after-training conditions. After training, however, there was less activity compared to the before-training condition. No brain area was specifically more activated in the automatic stage. These findings are similar to the results of young subjects and support our previous observation that no additional areas are activated specifically for automaticity in a self-initiated memorized sequential movement (Wu et al. 2004).

To perform sequential movements automatically, aged subjects had more activity than young subjects in the bilateral cerebellum, bilateral premotor area, bilateral parietal cortex, caudate nucleus, pre-SMA, anterior cingulate motor cortex, thalamus, left dorsal lateral prefrontal cortex and occipital cortex; and less activation in the left (contralateral) SM1. Our finding of increased neural network activation for aged subjects is similar to previous observations on age-related changes in cognitive and motor circuitry (Sailer et al. 2000; Calautti et al. 2001; Hutchinson et al. 2002; Mattay et al. 2002; Ward & Frackowiak, 2003). In our study, greater activation in aged subjects was found in more brain areas compared to those studies, which must be attributed to the more complex nature of our protocol. It is well known that more complex movements require greater brain activation (Sadato et al. 1996; Catalan et al. 1998; Wu et al. 2004). A previous study on motor sequence learning (Daselaar et al. 2003) found no age-related difference. The possible reason for this difference from our study might be ascribed to the use of different protocols. In the present study, we used self-initiated, self-paced, memorized sequence movements, whereas Daselaar and colleagues used an externally triggered, implicitly learned motor task.

There are several factors that may account for the observed age-related differences while performing automatic movements. One is the difference in the task performance. In our study, accuracy and rate at the automatic stage were the same for both groups. We did not measure other features, such as strength and velocity, which are known to decline with ageing (Smith et al. 1999). Previous studies have shown that decreased force is associated with less brain activation in areas such as the contralateral SM1 and SMA (Dettmers et al. 1995; Cramer et al. 2002). We found less activity in the contralateral SM1 in aged subjects, perhaps due to less vigorous movement. However, there was no difference in activity in the contralateral (right) SM1 between the two groups in the control experiment (left hand tapping). Hutchinson et al. (2002) found similar results and suggested that the difference in SM1 could not be explained by the difference of strength. Most importantly, we found more activity in extensive areas in the aged subjects. Therefore, although a subtle difference in motor performance might exist, the differences noted between aged and young subjects during execution of automatic movements cannot be attributed to task performance.

Age-related differences in the haemodynamic coupling of signal change between the two groups may also contribute to the differences in the brain activity. Previous investigations found some decreases, but no increases in the blood oxygen level dependent (BOLD) signal response in aged subjects as a result of altered haemodynamic coupling (Ross et al. 1997; D'Esposito et al. 1999). It has been suggested that if, in aged subjects, less activation in some brain regions is accompanied by more activity in other regions, as in our study, it is unlikely that regional variations in the haemodynamic coupling of neural activity to imaging signal would account for the age-related differences in brain activation (D'Esposito et al. 1999).

The differences between the two groups during execution of automatic movements might also arise from a difference in baseline resting activity. Some positron emission tomography (PET) experiments showed that aged subjects compared to young subjects have lower resting brain regional cerebral blood flow and glucose consumption in prefrontal areas (Calautti et al. 2001). With the technique employed in the present study, we could not compare the resting neural activity between the groups. However, we found more ‘deactivation’ in the prefrontal and anterior cingulate areas in young subjects compared to aged subjects (Fig. 5). These more deactivated areas corresponded to some of the more activated areas in the aged group (Fig. 3). This suggests that the age-related difference in resting metabolic state and deactivation might influence our results. However, this influence was limited to a few areas and could not explain the significant differences in motor network activity between the groups.

Some other factors might also have influenced our results. For example, some medications may affect brain excitability. Because all of our aged subjects were healthy and not taking any centrally active medications, this potential factor could be excluded. Another possible factor is that aged subjects might exhibit heightened levels of anxiety when attempting to perform the tasks correctly, which might also contribute to the observed difference of brain activity between the groups. However, our aged subjects were well trained and all of them were scanned while performing tasks correctly without feeling any difficulty. Moreover, the reduced activity in some regions and increased activity in others can rule out a significant influence of this factor on our results.

If the differences in activation between the two groups could not be attributed to the above factors, then it should reflect a different strategy in the neural network to produce the same performance. We found less activation of the contralateral SM1 in aged subjects. This contrasts with most previous studies, in which more activity in the contralateral SM1 was found in aged subjects (Mattay et al. 2002; Ward & Frackowiak, 2003). A possible reason for this difference might be the different protocols employed. Several studies on visual perception also showed decreased activity in the primary visual cortex in aged subjects (Ross et al. 1997; Grady, 2000). It was suggested that less activity is usually related to poorer performance, and aged subjects who can perform better have more activity (Cabeza et al. 2002; Mattay et al. 2002). This explanation is unsuitable for our data because there were no differences in performance between the groups. Because the SM1 is more involved in processing complex sequential movements than simple repetitive movements (Catalan et al. 1998), presumably, during performance of simple movements, aged subjects could increase the utilization of SM1 to maintain performance level (Mattay et al. 2002; Ward & Frackowiak, 2003). For complex movements, the activation in SM1 also significantly increased in young subjects, but in contrast, in aged subjects SM1 activity did not increase further. Therefore, SM1 in aged subjects was less activated compared to young subjects.

Among the areas additionally activated in aged subjects, the pre-SMA (Jenkins et al. 2000; Cunnington et al. 2002), anterior cingulate motor cortex (Frith et al. 1991; Jueptner et al. 1997a; Petersen et al. 1998), caudate nucleus (Alexander & Crutcher, 1990; Jueptner et al. 1997b), posterior parietal cortex (Deiber et al. 1996) and premotor cortex (Halsband et al. 1993) have a crucial role in planning and selecting a motor task. The dorsal lateral prefrontal and anterior cingulate cortices are important in monitoring task execution (Owen et al. 1996; Jueptner et al. 1997a). The premotor cortex, posterior parietal areas and cerebellum have greater activity as the complexity of the movement increases (Sadato et al. 1996; Catalan et al. 1998; Wu et al. 2004). In addition, the neocerebellum (Jenkins et al. 1994; Jueptner et al. 1997b), pre-SMA (Sakai et al. 1998; Hikosaka et al. 1999), caudate nucleus (Hikosaka et al. 1999; Nakano, 2000) and dorsal lateral prefrontal cortex (Jenkins et al. 1994; Jueptner et al. 1997a, b; Jansma et al. 2001) are important for acquiring new sequences. It was shown that aged subjects have greater difficulty in planning and sequencing a motor task than young subjects (Krampe, 2002). Therefore, our results suggest that aged subjects must recruit more areas to compensate for the greater difficulty they have in executing automatic movements. Even if there is no subjectively greater behavioural effort compared to the young subjects, their brain may work harder to perform automatically.

Further support for the idea that the demands differ between the two groups comes from within-group comparison in the before-and after-training stages. The differences in brain activity between the before-and after-training stage were suggested as being related to the process of automaticity (Wu et al. 2004). In the aged group, only the pre-SMA, premotor and parietal cortices were less activated at the automatic stage compared to the pretraining condition (Fig. 2). In contrast, in the young group, less activity was found in the cerebellum, premotor area, parietal cortex, pre-SMA, anterior cingulate cortex, caudate nucleus and dorsal lateral prefrontal cortex during the automatic stage (Wu et al. 2004). These results suggest that the pattern of brain activity of aged subjects in performing automatic movements is similar to that at the pretraining stage of young subjects and is obviously less efficient.

An additional factor that contributes to the observed age-related difference of brain activity is the timing-control of movements. Similar to our previous study (Wu et al. 2004), we did not use external cues to help subjects maintain the rates because the need for attention to follow the pace would weaken the claim for automaticity. Since the rate of movement has a significant effect on brain activity (Deiber et al. 1999), before the fMRI scan we gave all subjects sufficient time to practise the rate. Our results prove that aged subjects can execute simple and regular rates at the same level as young subjects (Krampe, 2002). However, aged subjects reported more difficulty than young subjects in achieving the required rate, which suggests that aged subjects have greater difficulty in timing-control (Krampe, 2002). Some brain areas, i.e. cerebellum, pre-SMA, dorsal lateral prefrontal cortex and basal ganglia, are involved in generating accurate timing of movement (Kawashima et al. 2000; Dreher & Grafman, 2002). The pre-SMA and dorsal lateral prefrontal cortex are especially important for self-paced movements (Wessel et al. 1995; Kawashima et al. 2000). Therefore, increased activity in these areas may be partly due to the additional brain effort aged subjects used for timing-control.

The difference in brain activation between the two groups may also be due to reorganization of a neural network in response to neurodegeneration. There is a series of changes during ageing, such as cell loss, synaptic degeneration, blood flow reduction and neurochemical alteration (Raz, 2000). Extensive studies have revealed central neural system reorganization following brain lesions, amputations and blindness (Cohen et al. 1997; Hallett, 2001). It is plausible to assume that some reorganization occurs in the ageing brain as a consequence of normal ageing (Buonomano & Merzenich, 1998).

Effect of ageing on dual task

Our knowledge of how normal ageing affects the execution of dual tasks is limited. In our study, at the before-training stage, performance of dual tasks by aged subjects was significantly worse than young subjects (Table 1). This result supports the previous finding that aged subjects perform dual tasks more poorly than younger subjects (Rogers et al. 1994; Rubichi et al. 1999; Li et al. 2001). After practice, although two aged subjects still had difficulty in performing dual tasks correctly, most of them could execute the dual tasks at the same level as young subjects. Our data demonstrate that the ability to perform dual tasks is still relatively intact in aged subjects. Aged subjects had more brain activity than young subjects in performing dual tasks, which demonstrated that their brains needed to work harder to perform the dual tasks at the same level as young subjects. The most remarkable difference between the two groups was that the bilateral precuneus was additionally activated in dual tasks in aged subjects. In contrast, in young subjects no additional area was activated in the dual tasks; all areas activated in the dual task were also activated by one or both of the component tasks. It is still controversial whether there is a central supervisor (D'Esposito et al. 1995) or not (Passingham, 1996) while performing dual tasks. It is not the concern of this paper to explore why there was no area specifically activated in our young subjects. The function of precuneus is poorly understood. The additionally activated precuneus suggests that the aged subjects may need more preparation (Astafiev et al. 2003), working memory (Callicott et al. 1999) and monitoring (Gusnard & Raichle, 2001) to execute dual tasks compared to young subjects. They must recruit more brain areas to compensate for their difficulty in executing dual tasks. In a previous study of the ageing effect on dual tasks, Smith et al. (2001) found that the left prefrontal cortex only activated in the dual task in aged subjects. However, in that study, aged subjects had significantly poorer performance and the prefrontal cortex also specifically activated in young subjects with poor performance. In contrast, in our study all aged and young subjects had good performance after training, which may explain why we did not find that the prefrontal cortex additionally activated in dual task execution in aged subjects.

In summary, we found that although they have more difficulty, after extensive practice, most healthy aged subjects could perform some complex motor tasks automatically as well as young subjects. Aged subjects had greater activity in the bilateral anterior lobe of cerebellum, premotor area, parietal cortex, left prefrontal cortex, anterior cingulate, caudate nucleus and thalamus, and recruited more areas, including the pre-SMA and the bilateral posterior lobe of cerebellum, compared to young subjects during performance of automatic movements. Our results suggest that aged subjects appear to require more brain network activity to perform automatically at the same level as young subjects, which contributes to their difficulty in achieving automaticity.