Conceived and designed the experiments: MR. Analyzed the data: SB. Wrote the paper: SB MR. Interpretation of the data: SB FR CK AJF TR MW MR. Critical manuscript revision: FR CK AJF TR MW.
SB has received honoraries for lectures from Novartis and support for scientific symposia by Shire. FR, CK, TR, AF, MW and MR report no conflicting interests. This does not alter the authors’ adherence to all the PLoS ONE policies on sharing data and materials.
Hyperactivity is one of the core symptoms in attention deficit hyperactivity disorder (ADHD). However, it remains unclear in which way the motor system itself and its development are affected by the disorder. Movement-related potentials (MRP) can separate different stages of movement execution, from the programming of a movement to motor post-processing and memory traces. Pre-movement MRP are absent or positive during early childhood and display a developmental increase of negativity.
We examined the influences of response-speed, an indicator of the level of attention, and stimulant medication on lateralized MRP in 16 children with combined type ADHD compared to 20 matched healthy controls.
We detected a significantly diminished lateralisation of MRP over the pre-motor and primary motor cortex during movement execution (initial motor potential peak, iMP) in patients with ADHD. Fast reactions (indicating increased visuo-motor attention) led to increased lateralized negativity during movement execution only in healthy controls, while in children with ADHD faster reaction times were associated with more positive amplitudes. Even though stimulant medication had some effect on attenuating group differences in lateralized MRP, this effect was insufficient to normalize lateralized iMP amplitudes.
A reduced focal (lateralized) motor cortex activation during the command to muscle contraction points towards an immature motor system and a maturation delay of the (pre-) motor cortex in children with ADHD. A delayed maturation of the neuronal circuitry, which involves primary motor cortex, may contribute to ADHD pathophysiology.
As hyperactivity is one of the core symptoms of attention deficit hyperactivity disorder (ADHD), it is crucial to understand the role of the motor system in the disease. Structural magnetic resonance imaging (MRI) studies and recent functional imaging data indicate that a disturbance of motor function in the primary motor cortex might contribute to ADHD pathophysiology
Indirectly, the top-down control of the motor system can be assessed when MRP are examined under specific conditions. An increased number of very slow responses in children with ADHD may reflect more frequent lapses of attention. Thus responses with long reaction times have been used to examine functional states in which the subjects are less concentrated than in trials with short reaction times
Finally, the motor system is crucially influenced by dopamine. MRP could reflect an excellent neurophysiological marker to monitor the effects of stimulant medication in ADHD
In order to characterize how ADHD, response speed (i.e. more or less concentrated states) and stimulant medication would differentially affect the neuronal activation related to triggering a movement (iMP) or its post-processing in short term motor memory (mPINV), we analysed lateralized MRP in a previously characterized sample of children with ADHD and matched controls
Seventeen boys with attention deficit hyperactivity disorder according to the ICD-10 criteria for F90.0 (corresponding to the DSM IV combined type of ADHD) by an interdisciplinary team (mean age 9.5±1.5 years; range 7.2 to 11.7 years; IQ 106±14.1; range 83–121) and 20 age- and IQ-matched healthy control boys (mean age 9.9±1.1 years; range 8.2 to 11.8 years; IQ 111±13.2; range 97–132) were recruited (group means were matched). For the response-locked EEG analysis in the current paper, we had to exclude one ADHD patient due to an insufficient number of successful response trials (N<10). The remaining sample consisted of N = 16 patients (mean age 9.6±1.5 years, IQ 104±14.3). IQ was assessed by the German version of the Wechsler Intelligence Scale for Children (HAWIK III).
Two reports about stimulus-locked data analysis in this sample have been published before
The study was approved by the ethics committee of the Medical Faculty of the University of Würzburg, Germany. All subjects and their parents provided written informed consent according to the Declaration of Helsinki.
Subjects performed a modified version of the continuous performance test (CPT-OX)
EEG was recorded from 21 gold cup electrodes which were placed according to the international 10–20 system using a sampling rate of 256 Hz and a band pass filter of 0.3 to 70 Hz. The current analysis focused on movement-related potential components which are evoked in response to cued reactions (details are given below). These components have a shorter duration
Only trials with correct responses within 1 second were included in the analysis. Data were segmented on response triggers from −2500 to 2000 ms. The first 200 ms of this epoch served as baseline. Taking into account that median reaction times were about 400–500 ms and that there was an interval of 1800 ms between the onset of the cue ‘O’ and the onset of the the target ‘X’, this baseline fell before the cue for nearly all responses in all subjects when fast reactions below the median were analyzed. We refrained from an even earlier baseline in order to avoid a contamination by preceding responses, though cues were preceded by distractors. We made sure that this baseline was not contaminated by lateralized responses to the cue or late MRP to the preceding trial. In
For control children, averages of all responses are illustrated together with a separate presentation of averages of fast (below median reaction time) and slow responses (above median reaction time). For children with ADHD, responses on and off methylphenidate are presented. For effects of response speed in ADHD see
Data were corrected for ocular artifacts using the algorithm according to Gratton and Coles (BrainVision Analyzer, BrainProducts, Munich, Germany). Artifacts were automatically rejected when the signal amplitude exceeded 150 µV due to the higher background EEG in children compared to adults. This procedure was controlled by visual inspection. The average reference was calculated offline.
Response locked lateralized MRP were assessed at C3 versus C4
Group differences in reaction times as well as effects of medication and learning were examined by Student’s t-tests.
Group effects: iMP’ and mPINV’ amplitudes were examined in an ANOVA with the between subject factor GROUP (unmedicated ADHD versus healthy controls) and the repeated measurement factor COMPONENT (iMP’ versus mPINV’). The factor component was introduced to test for differential group effects (interaction GROUP x COMPONENT) on pre- and post-movement potentials as these have been shown to differ in their maturational trajectories
Effects of stimulant medication (10 mg MPH): iMP’ and mPINV’ amplitudes were examined in an ANCOVA with the repeated measurement factors COMPONENT (iMP’ versus mPINV’) and MEDICATION (on/off methylphenidate). Again, median reaction times (off medication) served as a covariate in an additional analysis. Reaction time off medication was used as covariate instead of the mean between reaction times on and off medication because theoretically, medication effects on reaction time could have masked relevant findings. However, results did not change when the mean value of median reaction times on/off medication was used as a covariate instead of median reaction times off medication (not shown).
Finally, in order to assess the effects of response speed, a median split was performed into trials with slow and fast reaction times. An ANOVA with the between subject factor GROUP (ADHD versus healthy controls) as well as the repeated measurement factors COMPONENT (iMP’ versus mPINV’) and REACTION TIME (below versus above median reaction time) was calculated.
Pearson correlation coefficients between median reaction times and iMP’ amplitudes were calculated. This was done separately for the two diagnostic groups because we obtained group differences for iMP’ amplitudes (see group effects above).
Without medication, ADHD subjects had longer median reaction times compared to healthy control children (498±135 ms vs 421±77 ms; t = 2.2; p = 0.04). The difference was more pronounced for slow responses above (671±204 ms vs. 505±91 ms; t = 3.3; p = 0.002) than fast responses below median reaction time (405±79 ms vs. 359±69 ms; t = 1.9; p = 0.07). Median reaction time was reduced in ADHD subjects under methylphenidate compared to the run without medication (430±69 ms; t = 2.2; p = 0.04).
The time-course of potentials at central leads C3 versus C4 (
(top; from left to right: all responses, fast responses below median reaction time, slow responses above median reaction time)
ADHD | control children | |
mean ± standarderror (SE) | mean ± SE | |
iMP’ (without MPH) | 0.29±0.25 µV | −0.64±0.23 µV |
iMP’ (with MPH) | 0.03±0.33 µV | |
mPINV’ (without MPH) | −0.78±0.21 µV | −0.67±0.18 µV |
mPINV’ (with MPH) | −0.44±0.28 µV |
There was a main effect of COMPONENT as well as an interaction between diagnostic GROUP and COMPONENT in the ANOVA with these two factors (see
a) between-subject factor GROUP, repeated measurement factor COMPONENT | |||
GROUP | F(1;34) = 3.4 | p = 0.07 | |
COMPONENT | F(1;34) = 6.4 | p = 0.02 | |
GROUP x COMPONENT | F(1;34) = 5.6 | p = 0.02 | |
|
|||
iMP’: | GROUP | F(1;34) = 7.6 | p = 0.009 |
mPINV: | GROUP | F(1;34) = 0.1 | p = 0.72 |
Controlling for REACTION TIME as a covariate, the factors diagnostic GROUP and COMPONENT still interacted (F(1;33) = 4.9; p = 0.03). Newman Keuls post hoc tests showed that iMP’ amplitude was reduced in children with ADHD (p = 0.005; cf.
There were no significant differences in iMP’ amplitude between medication naïve children who were tested off medication first and children who had been medicated before and were tested on medication afterwards (t = 0.4; p = 0.68; unpaired t-test).
Stimulus locked waveforms are given for comparison in supplementary
The ANOVA with the factors MEDICATION and COMPONENT did not yield any significant main effect or interaction (
a) Repeated measurement factors COMPONENT and MEDICATION: | ||
COMPONENT | F(1;15) = 3.1 | p = 0.096 |
MEDICATION | F(1;15) = 0.0 | p = 0.90 |
COMPONENT × MEDICATION | F(1;15) = 2.6 | p = 0.14 |
|
||
off medication: | F(1;15) = 9.1 | p = 0.009 |
on medication: | F(1;15) = 0.7 | p = 0.42 |
Grand average of lateralized motor potentials ([C3
Response speed affected the group differences in MRP amplitudes (interaction GROUP x RESPONSE SPEED F(1;34) = 4.4; p = 0.04). Group differences were present only for fast reaction times below the median (p = 0.045) but not for slow reaction times above the median (p = 0.67; Newman Keuls post hoc tests;
Grand average of lateralized motor potentials ([C3
There was a trend towards a positive correlation between median reaction times and iMP’ amplitude in healthy controls (r = 0.43; t = 2.0; p = 0.06) but a negative correlation in the ADHD group (r = −0.50; t = 2.2; p = 0.048;
To our knowledge, this is the first study, which examines the influence of response speed and medication on movement-related potentials in ADHD, separating pre- and post-movement processing. Contralateral focal activation of the premotor and primary motor cortex during motor response programming (iMP’) exhibited a significantly reduced lateralization in children with ADHD compared to healthy controls, especially when reaction times were short. For the motor memory trace (mPINV’) this effect was not found, mPINV’ amplitude was even non-significantly higher (i.e. more negative) in trials with slow responses. Unlike control children, unmedicated children with ADHD showed significantly larger mPINV’ than iMP’ amplitudes. While in healthy controls shorter reaction times were associated with more negative iMP’ amplitudes, this pattern was reversed or at least absent in ADHD subjects. This finding appears plausible as several structural
Short reaction times may be taken as indication of better concentration on the task. Movement kinetics themselves (speed of the movement, muscle force) have been found to be largely independent of lateralized MRP amplitudes
Previous studies of lateralised ERP in ADHD support our findings and have shown a reduced contingent negative variation
Methylphenidate tended to normalize prolonged response latencies in ADHD children like in previous studies
Taken together, these findings indicate a qualitative difference in focal motor cortex activation in ADHD, which cannot be compensated for by medication or top-down control when only trials with short reaction times are taken into account. Previous studies have repeatedly shown a polarity reveral during childhood
The identification of lateralized movement-related potentials with their characteristic time course
We would like to emphasize, that ADHD is not a simple maturation delay, as findings about differences which persist into adulthood demonstrate. Some aspects of maturation seem to be delayed in ADHD and may contribute to (though not fully explain) ADHD pathology
A limitation of our study is the small sample size and that data from a continuous performance test were re-analyzed. The fact that subjects had to be prepared to inhibit their responses in some trials may have influenced our results. Future studies should include a standardized characterization of clinical motor problems in the examined sample and employ a wider range of motor paradigms ranging from freely selected spontaneous movements to pre-programmed movement sequences.
Response speed crucially modulates lateralized MRP amplitudes. Surprisingly, the most pronounced differences between ADHD and healthy control children were found in trials with fast reaction times, i.e. good concentration. The inverse association of response speed and iMP’ amplitude in the control and the ADHD group pointed towards a maturation delay in the motor system of ADHD children in our sample. Stimulant medication tended to normalize response speed, but did not normalize iMP’ amplitudes, giving further support to the hypothesis that the substitution of axodendritic by axosomatic synapses may be delayed in the motor cortex in ADHD children. This hypothesis warrants further investigation.
(TIF)
(DOC)
We thank Mr Benjamin Teufert, Dresden, for his help with the figure layout.