Unravelling developmental disregard in children with unilateral cerebral palsy by measuring event-related potentials during a simple and complex task

Cerebral Palsy (CP) is defined as a group of non-progressive disorders related to the development of movement and posture, caused by damage to the developing foetal or infant brain [1]. A large group (between 21 and 40%) of children with CP is formed by those with unilateral CP, having substantially greater motor deficits in one upper extremity than the other [2]. The observed unilateral motor impairments are related to damage of brain regions involved in planning, controlling, and execution of movements leading to a reduced movement capacity in children with unilateral CP [3]. Apart from the reduced movement capacity, a subset of children with unilateral CP also seem to disregard the preserved capacity of their affected upper limb, leading to a failure to use the affected arm and hand according to its full capacity in daily life [3‐6]. This discrepancy between capacity and performance is defined as Developmental Disregard (DD) [3‐5].

To date, different explanations have been put forward to explain DD in children with unilateral CP. A common explanation is based on the theory of operant conditioning [7, 8]. It compares DD to the phenomenon of learned non-use, defined as a learned suppression of movement, reported in the literature in adults who suffered a cerebrovascular accident (CVA) [9]. Following this theory, it is suggested that children with DD have experienced too little incentive to use the affected upper limb, because using the unaffected limb is less demanding [5]. Thus, positive reinforcement resulting from the successful use of the unaffected upper limb is combined with negative reinforcement from the unsuccessful use of the affected upper limb. This leads to a behavioural bias favouring the unaffected limb disproportional to the capacities of both the unaffected and affected upper limb.

Despite the similarity of the behavioural symptoms associated with learned non-use in CVA patients and DD in children with unilateral CP, recent studies emphasize that in DD both the developmental aspect and related cognitive aspects of information processing pose an important conceptual difference to the pure behavioural phenomenon described in learned non-use [3‐6, 10, 11]. In this respect, Deluca and colleagues [5] have postulated that children with DD have suffered a critical lack of movement stimulation during developmental periods when movement repertoires are rapidly acquired in typically developing children. As a consequence of this lack of movement that starts at perinatal periods, in combination with the earlier mentioned effects of reinforcement, typical developmental milestones are delayed or even deficient for the affected upper limb. In line, the neural substrates involved in motor control as well as in sensori-motor integration of the affected limb experience a similar lack in development and refinement [5, 10, 12]. It has even been stated that DD might be a neurologically based phenomenon similar to poststroke neglect syndrome [12].

In a recent explanation to account for DD this protracted development of motor control, sensori-motor integration and linked neural substrates is suggested to cause certain movement patterns of the affected arm and hand to be not sufficiently automated [4]. Based on Fitts and Posner’s [13] theory of motor skill acquisition, Houwink and colleagues [4] hypothesised that due to the lack of automaticity, using the affected upper limb requires a disproportional amount of attention [4]. They argue that a disproportional amount of attention coincides with an excess in cognitive load that is associated with motor control of the affected upper limb. The increased cognitive load in turn leads to a reduced spontaneous use of the affected arm and hand in daily life [4]. This hypothesis was already verified in several studies with CVA patients. These studies showed that patients, who have to relearn a lost motor skill, need a disproportional level of attention when moving the affected limb in the early stages of rehabilitation when relearned movements are not yet (re)automated [14, 15]. Thus, a lack of automaticity of movements is associated with increased cognitive load in adult CVA patients.

To be able to assess cognitive load related to movement, Event-Related Brain Potentials (ERPs) offer the unique opportunity to directly measure neural responses associated with distinct processing stages preceding an overt response [16]. Whereas mid-latency components (e.g. N1 & P2) have been associated with orienting and perception, the long-latency components of ERPs (e.g. N2, P3) are known to reflect processes associated with cognitive control and attention allocation [17, 18]. To assess the possible role of cognitive load in the impaired motor performance of DD, the current study therefore focussed on the long-latency N2b component. Next to generally being known to reflect processes associated with cognitive control and attention allocation, the N2b has also already directly been linked to cognitive control of response-related processes [19].

In order to investigate the aspects of information processing preceding goal directed motor responses, ERPs were extracted from the ongoing EEG during a single-hand task as an index of the individuals hand capacity and a dual-hand task, to estimate the hand performance. Based on the cognitive load theory of DD we reasoned that children with DD will show alterations linked to the higher order cognitive control processes when preparing a response with their affected upper limb during the dual-hand performance task. We therefore hypothesize that children with DD show alterations related to the N2b component when selecting a response with the affected upper limb, reflecting increased cognitive load in order to generate an adequate response. We furthermore hypothesize this effect to be especially pronounced during the more demanding dual-hand performance task.

To test our first hypothesis that children with DD compared to children with unilateral CP without DD show alterations related to long-latency ERP component when selecting a response with the affected upper limb compared to the non-affected upper limb, repeated measures GLM analyses of the long-latency N2b ERP component were performed. To ensure that differences can not be ascribed to earlier processes related to the evaluation of the physical features of task relevant stimuli, the mid-latency N1 and P2 components were also inspected. To furthermore ensure that differences could also not be ascribed to early cue evaluation or general visual stimulus evaluation processes, ERP components following cue stimuli were investigated as well.

The repeated measures GLM analyses of the ERP components following cue stimuli revealed no significant interaction or main effects with respect to the N1 component (all p’s > .10), the P2 component (all p’s > .10), or the N2b component (all p’s > .10). The analyses of the ERP components following target evaluation revealed no significant interaction or main effects with respect to the mid-latency N1 (all p’s > .10) and P2 components (all p’s > .10). With respect to the long-latency N2b component this analysis did however reveal a significant task (single vs. dual task; F(1,13) = 5.265, p = .041, η_p ² = .288) effect as well as a significant handedness x task x group interaction (F(1, 13) = 6.649, p = .026, η_p ² = .338).

To examine this interaction and to test our second hypothesis, that the long-latency effect is especially pronounced during the more demanding dual-hand performance task paired sample t-tests between hands (affected vs. non- affected hand) were performed for each group and task separately. This revealed that the significant difference between hands was only present in the DD group and only in the dual-hand task (t(6) = 2.469, p < .05). Specifically, only in the dual-hand task the N2b amplitude following target stimuli was significantly enhanced in the DD group when using the affected hand compared to using the non-affected hand, confirming our first and second hypothesis. Grand average ERPs to target stimuli for both the DD and the control group are depicted in Figure 2. Figure 3 provides a visual presentation of the mean absolute amplitude for the N2b component following target stimuli. No differences between hands were observed in the control group.

Finally, with respect to the behavioural data the repeated measures GLM analyses of the IES scores revealed a significant main effect of hand (F(1, 13) = 5.715, p = .033, η_p ² = .305). Across both group responses were less efficient with the affected hand (M = 598.17, SD = 256.48) compared to responses with the non-affected hand (M = 558.38, SD = 227.28). No GLM interaction effects with respect to the IES scores were observed. The IES scores were similar for both groups.

The goal of the current study was to use Event-Related Potentials (ERPs) to provide a direct measure of cognitive load associated with movements of the affected upper limb in children with unilateral Cerebral Palsy (CP) and Developmental Disregard (DD). Based on Fitts and Posner’s [13] theory of motor skill acquisition, it has been suggested that due to a lack of automaticity of movements with the affected upper limb, children with DD experience increased cognitive load when using this limb [4]. This increased cognitive load in turn leads to an underuse of the affected arm and hand in daily life performance even if sufficient limb capacity is available. In order to test this theory we recorded ERPs during two different versions of a cue-target paradigm. First, we employed a single-hand task as an index of the individuals hand capacity. Next, we recorded a dual-hand task to estimate the hand performance. We first of all hypothesized that children with DD would show alterations related to the long-latency ERP components when selecting a response with the affected upper limb, reflecting increased cognitive load when generating an adequate response. Secondly, we hypothesised this effect to be especially pronounced during the dual-hand performance condition, reflecting the characteristic discrepancy between hand capacity and hand performance of DD.

In line with our first hypothesis children with DD showed an enhancement in mean amplitude of the long-latency N2b ERP component when preparing a response with their affected compared to their non-affected hand. This component is well known to reflect the amount of activity in areas associated with cognitive load, indexes voluntary attentional processing, and is known to represent a necessary part of the information processing sequence leading to a motor response [30, 31]. The finding of the increased N2b following target stimuli in children with DD therefore indicates that these children experience an increase in cognitive load when generating an adequate motor response with their affected compared to their non-affected hand. Moreover, and in line with our second hypothesis, this enhancement was only observed in the dual-hand performance task and not in the single-hand capacity task.

Based on our electrophysiological results the conclusion is warranted that the discrepancy between capacity and performance in children with DD can be explained by an increased cognitive load associated with response selection in dual-hand task situations. This conclusion is further strengthened by three findings of the present study. First, the enhancement of the N2b ERP component following target stimuli when generating a response with the affected hand was not found in the control group. Second, we did not find any differences between hands with respect to the mid-latency N1 and P2 amplitude. This indicates that there are no differences between both sides of target presentation regarding the evaluation of the physical features of task relevant stimuli [32, 33]. Observed group differences can therefore not be explained due to differences in processes related to orienting and perception that regularly accompanies CP [34]. Third, and finally, there were no differences regarding cue evaluation processes preceding the target stimuli. This finding shows that visual and cognitive evaluation processes that are not directly linked to preparing a motor response following target evaluation are not impaired in children with DD compared to children with unilateral CP without DD.

The fact that preparing a response with the affected upper limb compared to preparing a response with the non-affected arm and hand increases cognitive load in a dual-hand performance task is in line with the behavioural observation of the discrepancy between capacity and performance that characterizes DD [4]. The current study showed that children with DD use their affected hand during a single-hand capacity task without any electrophysiological or behavioural indications of increased cognitive load when preparing the response. This finding is in agreement with the behavioural observation that children with DD are indeed able to perform a particular task with their affected hand alone, as long as they can focus on the task and their hand capacity is sufficient [4, 10]. However, we did observe an increased ERP N2b component when preparing a response with the affected upper limb in a dual-hand performance task in children with DD reflecting increased cognitive load. This finding exemplifies the behavioural observation that in spontaneous daily use, predominantly requiring both hands, children with DD fail to use the potential motor functions of their affected limb and rather chose to perform a task with their non-affected upper limb alone [4, 10]. In this connection, an interesting facet of the current study regarding the behavioural observations of DD is that there were no differences between groups regarding the behavioural efficiency scores. That is, even though our electrophysiological findings indicate increased cognitive load associated with the use of the affected upper limb in a dual-hand task an appropriate movement outcome could be achieved. These findings further substantiate the claim that DD is due to increased cognitive load associated with the use of the affected upper limb that only reveals itself in complex activities where attention cannot be solely focused on the effected arm and hand. Children with DD are able to use the affected hand efficiently even in dual-hand tasks but due to the enhancement of cognitive load associated with this movements they disregard their hand in spontaneous daily live.

In sum, the results of the current study add to the accumulating evidence suggesting that cognitive aspects of information processing play a major role in the appearance of DD [3‐6, 10, 11]. Furthermore, these results are in line with the assumption that DD might be a neurologically based phenomenon similar to poststroke neglect syndrom [12]. It is already known that motor neglect becomes worse when attention is distracted and that simultaneous movement of the opposite limbs may also increase motor neglect [35]. This comparison would also be in line with the theory that due to an asymmetrical development of the affected limb, new neural substrates for entire classes of behavior are not well established, refined, and coordinated [5, 10, 12].

Next to making a substantial step in unravelling DD in children with unilateral CP, these results have important clinical implications. To date, a very commonly applied therapy aimed at improving the upper limb capacity in all children with unilateral CP is the so called ‘forced-use’ or ‘Constraint Induced Movement Therapy’ (CIMT) [36]. The main characteristic of this therapy is the immobilization of the non-affected upper limb, thus forcing the patient to use the affected limb exclusively [36]. By applying this therapy the capacity of the effected arm and hand is intensively trained and often improves spectacularly [37‐40]. However, CIMT was originally developed to overcome learned non-use in adult CVA patients and to promote use of the limb rather than skill [36]. In children, however, compared to adult patients developmental factors play a major role in the occurrence of DD [3‐6, 10, 11] whereby they may have never learned how to effectively use their more affected upper limb during many tasks [6, 41]. Gordon [41] therefore concluded that treatments for children to overcome DD must be developmentally focused and must take into account the importance of motor learning. This critical view on applying CIMT to children with DD is strengthened by the findings of the current study. We showed that the performance issues typically observed in DD are directly related to increased cognitive load only when using the affected hand in a dual-hand performance task and not when simply moving that limb in a single-hand task. It should therefore be considered to apply bimanual training instead of CIMT to children with DD to promote bimanual skill instead of unimanual use.

Recently, CIMT treatments have been combined with, or compared to, bimanual training therapies [40‐45]. These studies show that the results of bimanual training therapies as well as a combination with CIMT are very beneficial and lead to similar improvements in hand capacity as CIMT. Furthermore, they showed that bimanual training leads to a further improvement in bimanual skill and self-determined goals.

Next to the promising consideration to apply bimanual training therapies to children with DD it should also be considered that upper-limb training should not end after an intensive rehabilitation program, but to be continued and integrated in the daily-live activities of the children with DD [4]. As been shown in adult CVA patients increased cognitive load is directly associated with a lack of automaticity of using the affected upper limb [14, 15]. In the current study we demonstrated that during hand movements of the affected hand in a dual-hand performance task a disproportional amount of cognitive load is activated suggesting a lack of automaticity of using that hand only during dual-hand performance tasks. Continuing and integrating the rehabilitation program in the daily-live activities of the children with DD therefore has to be considered to promote further automatisation of movements of the affected hand during daily dual-hand performance. In this respect it has already been reported that concerning the daily performance of the affected arm and hand in children with unilateral CP treatment is more effective when conducted in the home setting of the children compared to the clinical setting [46]. In the study of Rostami and Malamiri [46] it was concluded that this natural daily-life environment provides more information about upper limb performance than other contexts such as the clinic. Considering continuing training in a home setting after completing a bimanual rehabilitation program is therefore a crucial next step in reducing DD.

Study limitations include small sample size and therefore difficulties controlling for any interaction between gender and maturation. A further limitation of the current study, also related to the small sample size, is the heterogeneity of the studied group. This latter limitation is however inherent to the participant population as unilateral Cerebral Palsy comprises a very heterogeneous group of movement disorders.

The discrepancy between capacity and performance in children with DD can be explained by an increased cognitive load associated with response selection in dual-hand task situations. The results of the current study therefore provide direct neurophysiological evidence to the accumulating indications suggesting that cognitive aspects of information processing play a major role in the appearance of DD. Furthermore, by showing that the performance issues typically observed in DD are directly related to increased cognitive load only when using the affected hand in a dual-hand performance task and not when simply moving that limb in a single-hand task it can be concluded that bimanual training, instead of CIMT, should be applied as therapy to children with DD.