Delayed development of visual motion processing in childhood migraine

Introduction

Migraine is a chronic, predominantly genetically determined disorder, characterized by recurrent attacks of headache, associated with features of autonomic disturbance. While probably not so well recognized, childhood migraine is also a widespread problem. Alterations of the visual system have already been demonstrated in migraine (1). In particular, evidence suggests that motion perception, a function mostly associated with the occipitoparietal visual system (or dorsal stream), is also affected (2–4). Data on similar alterations of visual perception in paediatric migraine are scarce (5). However, we have previously demonstrated magnocellular/dorsal stream deficits in children suffering from migraine without aura, similar to those found in adult migraineurs (6).

In a random dot kinematogram paradigm, Antal et al. (4) found deficiency of coherent motion detection in adult migraineurs, which was explained in terms of increased cortical excitability. This explanation was based on the observation that externally induced increased cortical excitability in human V5/MT in healthy individuals resulted in motion detection threshold increase, as measured with the help of a random dot kinematogram (7). It was suggested that increased cortical excitability leads to increased noise, which makes it more difficult to identify coherent motion. That is, the observed phenomenon was explained in terms of altered filtering ability.

In the present study we set up a random dot kinematogram-based motion transparency paradigm in order to compare motion coherence thresholds in paediatric migraine without aura with those measured in healthy controls, and to see if the deficit described in adult migraineurs can already be found in paediatric migraine.

Methods

The 14 migraineurs (without aura), aged between 8 and 17 years, were all patients of our outpatient clinic. The diagnosis was based on the criteria of the International Headache Society (8). Patients had only taken over-the-counter painkillers before initiation of the study. Only migraineurs with no other known neurological or psychiatric condition were included in the study. The 21 controls (of the same age range) were recruited. All subjects had normal or corrected-to-normal (20/20) visual acuity. The most important migraine-related and demographic characteristics of the studied population are summarized in Table 1. Testing was done interictally in all cases. The study was approved by the Ethics Committee of the University of Szeged, and conformed to the tenets of the Declaration of Helsinki in all respects. To testify their informed consent to the participation of their children, parents were asked to sign an informed consent form.

OPEN IN VIEWER

Table 1.

Characteristics of migraineurs and controls

	Number of participants (n)	Mean age (years)	Gender ratio (nF/nM)	Disease duration (years, average)	Days elapsed since last attack on the day of measurement (days, mode)	Positive family history (n)	Number of attacks/month (mode/range)	Unilateral location (% (n))	Pulsating quality (% (n))	Severity: severe* (% (n))	Aggravated by routine exercise (% (n))	Nausea / vomiting (% (n))	Photophobia / phonophobia in headache phase (% (n))
Migraine	14	13 (range: 8–17)	7/7	2, 13	4	8	2 (1–4)	71 (10)	100 (14)	86 (12)	71 (10)	43 (6)	57 (8)
Control	21	12, 4 (range: 8–17)	10/11	NA	NA	3	NA	NA	NA	NA	NA	NA	NA

Stimuli were generated with Psychophysics Toolbox Version 3 (http://psychtoolbox.org/), under MatLab (MathWorks, Inc.) on a PC, and presented on a 24-inch LCD monitor at a resolution of 1920 by 1200 pixels and at a 60 Hz refresh rate. Stimuli were random dot kinematograms with variable coherence rates (Figure 1). Stimuli consisting of 100 moving dots were presented on a neutral grey background in a centred rectangular stimulation field occupying 60% of the whole screen. At a viewing distance of 0.5 m (as during the trials), the stimulation field subtended an area of 35.74° by 22.34°. The diameter of each dot was 10 pixels (∼3 mm), subtending ∼0.34°. In each trial a given percentage of the dots moved coherently to the right or to the left, while the rest moved in random directions. After each trial, movement starting points were regenerated so that subjects would avoid using the movement of one dot as a clue. One trial lasted approximately 0.8 s (50 consecutive frames), during which each dot travelled 38.4 mm at a speed of 48 mm/s. Therefore, the path of movement subtended a visual angle of 4.4° for each dot. OPEN IN VIEWER

The task of the subjects was to indicate whether the coherently moving dots moved to the left or to the right by pressing the appropriate cursor button on a computer keyboard. The absolute coherence threshold was determined by QUEST adaptive threshold seeking algorithm (9) (Figure 2). OPEN IN VIEWER

For the statistical analysis we used Statistica for Windows 9 (StatSoft, Inc.). Because of the relatively low number of participants, we did not divide the groups into cohorts by age. However, we made sure that the demographic characteristics of migraineurs and controls would be as similar as possible (see Table 1), so that the comparisons would be valid. Due to the lack of normal distribution in both the migraineur group (Shapiro–Wilk W = 0.94, p = 0.44) and the control group (Shapiro–Wilk W = 0.92, p = 0.07), we chose the non-parametric Mann–Whitney U-test for the comparisons. Correlations (Spearman’s R) between task performance and certain migraine characteristics (like positive family history and attack frequency) were also calculated.

Results

The comparison of the two groups revealed a significant difference in motion coherence detection threshold, the control group having a lower threshold (MWU = 62.5, N₁ = 14, N₂ = 21, p < 0.05, two-tailed; healthy: median (range) 0.2 (0.18–0.23), migraineurs: median (range) 0.32 (0.14–0.56)). The difference between the two groups seems to be more pronounced for girls (MWU = 14.5, N₁ = 10, N₂ = 7, p < 0.05, two-tailed) than for boys (MWU = 18.5, N₁ = 7, N₂ = 11, p = 0.07, two-tailed). No significant gender-related difference in performance was found within the groups: (MWU = 18.5, N₁ = 7, N₂ = 011, p = 0.07, two-tailed, migraineurs) and (MWU = 41.5, N₁ = 10, N₂ = 11, p = 0.35, two-tailed, controls).

A more detailed analysis of the age dependence of the motion detection threshold revealed that the control group performed at a constant level, regardless of age, while the motion coherence detection threshold of migraineurs was higher at younger ages, catching up with controls by late puberty (Figure 2). In statistical terms: regression coefficients were significantly different (β_migraine = –0.3; β_controls = −0.001, t = –5.68, p < 0.001). This was also reflected in correlation coefficients (threshold vs. age): R = 0.685, p < 0.05, migraineurs; R = 0.064, n.s., controls.

A further finding is that migraineurs with direct line relatives suffering from migraine seem to exhibit poorer performance: (MWU = 6.0, N₁ = 6, N₂ = 8, p < 0.05, two-tailed). The length of disease history or monthly occurrence did not exhibit significant correlation with performance: R = 0.46, n.s.; and R = 0.27, n.s., respectively.

Discussion

In the present study we have found reduced motion coherence processing capacity in children suffering from migraine. We have also demonstrated a delayed development of visual motion processing in the age range of 8–17 years. To understand possible mechanisms, it is interesting to consider neuronal development early in life.

Post-mortem examinations showed that during the first 2 years of postnatal development, there is an excessive overproduction of synapses, followed by a longer period of activity-dependent pruning of synaptic connections (10). This pruning ends by mid-adolescence in the motor system, but it is accomplished earlier in the visual system (11). Another ongoing maturation process during childhood and adolescence is myelination (12). Kanemura et al. (13), in a volumetric MRI study, found that prefrontal cortex reaches its final, adult-like size by the age of 18, with a period of rapid growth between the ages of 8 and 14. Furthermore, Fornari (14) found that there is a linear relationship between age and white matter volume between the ages of 7 and 13, which is also reflected by improving performance on a spatial integration task. Olesen et al. (15), in a combined DTI-fMRI study, showed that the myelination of long-range cortical connections and the activation of their target areas are associated between the ages of 8 and 17. All these findings reflect a prolonged period of increased developmental vulnerability.

A possible explanation for the delayed development involves the noxious changes that occur in the diseased brain, and that might interfere with normal development. A likely candidate is cortical spreading depression (16) that might trigger, or in itself consist of noxious processes (17) sufficient to cause neuroinflammation, pain (18–20), and possibly cellular damage. The excessive release of glutamate (21) is also considered to induce excitotoxicity and cell death (22). An increased level of matrix metalloproteinase activity has recently been detected in migraineurs (23), possibly bringing about the leakage of the blood–brain barrier and also leading to an inflammatory response and neuronal damage (24). Similarly, Yilmaz et al. (25) found increased ictal levels of S100B (a marker of glial damage) and neuron-specific enolase (a marker of neuronal damage) in migraineurs without aura.

In an earlier study, Antal et al. (4) have shown similarly altered motion detection in adult migraineurs. The underlying mechanism was suggested to be the altered noise filtering caused by increased cortical excitability (26). In children, a recent TMS study also found reduced phosphene thresholds in the interictal, as well as in the pre- and post-headache periods, which also corroborates the presence of increased visual cortical excitability in young migraineurs (27). However, it must be noted here that a recent study by Webster et al. (3) came to the conclusion that the motion coherence processing alterations in migraine are independent of noise filtering.

We also have to consider that changes causing the altered motion processing in children with migraine might not directly be found in V5/MT. It was demonstrated that the processing of special features of visual stimuli can be modulated by attention (28). Similarly, primate studies (29) and fMRI investigations (30,31) showed attentional modulation of motion processing. Therefore, the developmental difference we observed in our study may also be associated with the delayed/altered maturation of modulatory connections.

Because during early development the maturation of the nervous system is activity-dependent (10), it can be hypothesized that the altered cortical excitability and the repetitive painful episodes might modify normal development in children by a mechanism of maladaptive plasticity. This would considerably slow down maturation, potentially resulting in a delayed development of visual function.

In summary, our results demonstrate a special feature of childhood migraine, which – after further in-depth investigations – might be used as an easy-to-use, non-invasive biomarker to monitor disease progression. Furthermore, corroborating our previous studies on visual processing in paediatric migraine (6,32), these results point to an altered central nervous system development, which implies the necessity of further studies on various sensory and cognitive functions in this condition.