Diagnostics and Treatment of Dyslexia

Aims: QEEG and neuropsychological tests were used to investigate the underlying neural processes in dyslexia.

Methods: A group of dyslexic children were compared with a matched control group from the Brain Resource International Database on measures of cognition and brain function (QEEG and coherence).

Results: The dyslexic group showed increased slow activity (Delta and Theta) in the frontal and right temporal regions of the brain. Beta-1 was specifically increased at F7. EEG coherence was increased in the frontal, central and temporal regions for all frequency bands. There was a symmetric increase in coherence for the lower frequency bands (Delta and Theta) and a specific right-temporocentral increase in coherence for the higher frequency bands (Alpha and Beta). Significant correlations were observed between subtests such as Rapid Naming Letters, Articulation, Spelling and Phoneme Deletion and QEEG coherence profiles.

Discussion: The results support the double-deficit theory of dyslexia and demonstrate that the differences between the dyslexia and control group might reflect compensatory mechanisms.
Integrative Significance: These findings point to a potential compensatory mechanism of brain function in dyslexia and helps to separate real dysfunction in dyslexia from acquired compensatory mechanisms.

Keywords: Dyslexia; EEG; QEEG; coherence; double-deficit theory

Acknowledgements
Data from The Brain Resource International Database was generously provided by the Brain Resource Company Pty Ltd. We would also like to thank local BRC clinics for data collection of the control group. All scientific decisions are made independent of Brain Resource Companies commercial decisions via the independently operated scientific division - BRAINnet - which is overseen by the independently funded Brain Dynamics Centre and scientist members. We would also like to acknowledge the contributions from Ine Giepmans and Minnie War for the data collection and Sabine de Ridder, Aukje Bootsma and Hanneke Friesen for acquiring the QEEG and neuropsychological data.

Introduction
Developmental dyslexia is characterized by difficulties with accurate and/or fluent word recognition and by poor spelling and decoding abilities. These difficulties typically result from a deficit in the phonological component of language that is often unrelated to other cognitive abilities [20]. Dyslexia is probably the most common neurobiological disorder affecting children, with prevalence rates ranging from 5 to 10 percent. It is a persistent, chronic condition [33].

     Our aim was to compare brain function of dyslexic children with non-dyslexic children on different neurophysiological and neuropsychological measures. Our question focussed on whether different QEEG activation patterns can be found in dyslexia and to what extent correlations between reading and spelling abilities and specific tasks for rapid naming and phonological awareness can be found to address the double-deficit theory of dyslexia [43]. We also assessed neuropsychological function in these groups in order to exclude further cognitive differences between the groups potentially confounding the QEEG findings. Our hypothesis was that the groups are not different on neuropsychological tests and that children with dyslexia show increased inter- and intrahemispheric coherence.

Results
EEG Power
An overview of Delta and Theta power for all sites is depicted in Figure 1 and the significant differences are indicated. The following differences between the two groups were found:
Delta: increased Delta power for the dyslexia group at Fp1 (F=6.315, df=1,33 p=.017), Fp2 (F=4.861, df=1,34 p=.034), F7 (F=4.806, df=1,34 p=.035) and T6 (F=6.193, df=1,35 p=.018).
Theta: increased Theta at Fp1 (F=11.072, df=1,33 p=.002), Fp2 (F=5.074, df=1,34 p=.031) and F7 (F=8.267, df=1,34 p=.007).
Beta 1: increased beta-1 at F7 (F=4.450, df=1,34 p=.042).

Fig. 1.      Mean EEG power for Delta and Theta for the dyslectic group (in red) and the control group (in black). All findings have p<0.05.

QEEG Coherence data
Figure 2 gives an overview of the significant increases in coherence per frequency band. All increases have p values of p<0.001. Due to the many significances no detailed statistics are given. All coherence values were increased for the dyslexia group red connections are from homologous pairs (both right and left hemisphere) and purple are uniquely right- or left hemispheric increased coherences.  Note the specific patterns which include mainly frontal, central and temporal sites. Also note the bi-laterally increased delta coherence fronto-central and the right fronto-central increased coherence specifically in the alpha and beta band.

Fig. 2 Significant increased coherences for the dyslexic vs. the control group for the different frequency bands. Red connections are from homologous pairs (equal right and left) and purple are unique right- or left hemispheric increased coherences. All p’s<0.001.

Neuropsychology
The dyslexia group named fewer words in the word condition of the Verbal Interference test (F=-6.994, df=36, p=.012) but there was no difference on the color condition of the (F=2.330, df=36, p=.136). The dyslexia group recognized fewer words as compared to the control group (F=8.914, df=36, p=.005) on the memory recognition task.

Within group correlations.
The within group correlations were performed on 18 dyslexic children only, one subject was removed from the analysis due to his age. This child was 16 year old whereas the majority of the group was around 10 years of age. Including him could have lead to spurious age related correlations.
Figure 3 shows the significant correlations between the obtained significant measures reported in the previous section and the sub-tests used to measure the severity of dyslexia. All significant EEG power and EEG coherence measures (63 measures: 8 EEG and 55 coherence) were submitted to the correlation analysis with the  4 dyslexia sub-tests: Rapid Naming Letters - RNL; Phoneme Deletion - PD, Articulation – ART and Spelling SPL. The results are depicted in 4 different colors, each depicting a significant correlation between that variable, between those locations for the given frequency band. The thickness of the line also depicts the significance level (thin p<0.05; thick p<0.001).
Interestingly, there were no significant correlations between the EEG power data and the EEG Coherence data within frequency bands, hence the increased coherence for dyslexic children cannot be explained by the increased delta and theta frontally.
 There was only 1 significant correlation, between EEG power and the severity of Dyslexia: the power of Theta at FP1 and Spelling (r=.510; df=16; p=.044).
For coherence the significant differences are depicted in Table 1 and are also visually depicted in Figure 3:

Fig. 3.      Significant correlations between dyslexia sub-scales and significant EEG Coherences in the different EEG bands. Red = Delta Coherence; Orange = Theta Coherence;  Light Blue = Alpha Coherence; Dark Blue = Beta Coherence;. Thick lines represent significances of p<.001 and thinner lines of p<.05. Note the specific independent patterns for some of the patterns especially Articulation with a centro-temporal pattern but also the continuous involvement of the right temporal region for all measures. Also note the clear differences between  the slow (Delta and Theta; Red and Orange) vs. higher (Alpha and Beta; Light and Dark Blue) EEG frequencies and the similarities between Delta/Theta and Alpha/Beta.

Discussion

This study focused on brain function patterns and neuropsychological findings in children with developmental dyslexia and tried to establish a relation between QEEG parameters with dyslexia relevant constructs. EEG findings show an increased (left)frontal and right temporal slow activity in the Delta and Theta bands and increased Beta 1 power at F7. Since all EEG data have been EOG corrected using Gratton et al [11] it is very unlikely the frontal increased Delta and Theta is due to residual EOG. QEEG Coherence data showed increased coherence in frontal, central and temporal regions. However, the increased coherences seemed to show a frequency specific effect, where the slower frequencies (delta and theta) showed a more symmetrical increase in right and left frontal, central and temporal networks whereas the higher frequencies (alpha and beta) showed a more specific right-hemispheric effect originating at T4 and F8. Correlational analysis showed that these increased coherences were an effect in itself, since there were no correlations between the increased delta and theta power on the one hand and the increased coherence in the according band on the other hand; hence the increased delta and theta power are not the cause of the increased coherences. High coherence between two EEG signals means high cooperation and synchronization between underlying brain regions within a certain frequency band [39]. Increased coherence can thus be interpreted as increased functional connectivity. This could implicate that dyslexic children have more activated frontal, central and temporal networks.

There were also significant differences between the dyslexic and the control group on the verbal interference tests (similar to the Stroop test) and the memory recognition test. These findings are directly related to the dyslexic problems as dyslexic children have decoding problems. The dyslexic children named less words on this test than the control children but had no impairments with respect to the colour condition. Dyslexic children also recognized fewer words on a memory recognition tests whereas spontaneous memory recall was not affected at all. These findings implicate that interpretation of performance on neuropsychological tests on these specific tests should always be treated with caution and maybe that dyslexic status should be incorporated into the interpretation of neuropsychological data to safeguard false positive findings on these tests.

Correlational analyses showed a correlation between the obtained significant QEEG findings and the  tests: articulation, rapid naming of letters, spelling and phoneme deletion. The correlated patterns – as depicted in Figure 3 – show quite specific patterns for all these 4 sub-tests. Interestingly, all these correlations were positive and quite high (explaining > 30% of the variance) meaning that better performance on these tests was associated with increased coherence. Given the fact that all these coherences were increased in comparison to the control group, one might conclude that these patterns reflect compensatory mechanisms and do not explain the deficit per se since then negative correlations were expected. The EEG in this study was NOT recorded during the dyslexia specific tests hence we recorded passive eyes open EEG which correlated highly with these tests. This further supports the fact that these patterns can be considered compensatory patterns since they are also present at rest. Furthermore, this also demonstrates that clear associations can be found between passive brain states and deviant behaviour demonstrating the utility of integrative approaches.
There seems to be a clear distinction between delta coherence on the one hand and beta coherence on the other. The increased coherence for dyslectic children was prominent and symmetric for the delta band; but rather local at the left side for the beta band. The correlations also further demonstrate this; a slow (delta & theta) coherent network left frontal and central and a faster (alpha & beta) network originating at T4. Although EEG coherence between different cortical regions is largely established by cortico-cortical and thalamo-cortical interactions [24] subcortical brain areas contribute to both inter- and intrahemispheric functional communication as well [3]. Especially the lower bandwidths such as the delta frequency in the EEG coherence spectrum have been associated with limbic contributions to cortico-cortical coupling [18], hence for this increased low-frequency coherences a limbic contribution could be hypothesized.

The core dysfunction in dyslexia hence seems to consist of increased slow activity left frontal and right temporal (T6) and bilateral increased coherence in the slower frequency bands (Delta and theta) as opposed to acquired-compensatory mechanisms consisting of right-hemispheric increased coherences in the higher frequency bands (alpha and beta) and a left frontal increased coherence in slower bands originating from C3 and FC3. The numerous increased coherences in the delta band fronto-central tend to suggest a strong limbic involvement as part of the core deficit in dyslexia, but this requires further study.
In this study, children showed indeed delays in both rapid naming and phonological awareness. This seems to correlate with the activation in the frontal-cerebellar phonological system. The QEEG findings in our study showed an increased activation pattern in dyslexic children, mainly in frontal and temporal lobes. Furthermore, the correlational analyses showed significant correlations with spelling, phonological skills and rapid naming with quite different topographical representations, suggesting involvement of different neural mechanisms. It can tentatively be concluded that the frontal-cerebellar network may be critical to the precise timing mechanisms that underlie the double-deficit theory of dyslexia, suggesting the existence of three subtypes of reading disability: dyslexics with deficiencies in phonological skills, poor rapid naming skills or a combination of both types. Thus, the present study supports the theory of Eckart et al. [4] hypothesizing that impairments in a frontal-cerebellar network may play a role in delayed reading impairment in dyslexia. These authors reported that anomalies in a cerebellar-frontal circuit are associated with rapid automatic naming and phonological processing.

Past EEG studies have shown different findings. Rippon & Brunswick [28] found no specific activation patterns with respect to dyslexic children. Weiss & Mueller [39] have proposed several roles for coherence in the different frequency bands (also see introduction) however in this study, we did not use a task-related protocol, that is, our EEG recording and analysis is restricted to a passive eyes open task, making comparison to her study very difficult.
This study contributes to the theory that neurobiological causes underlie dyslexia. The increased activation patterns of dyslexic children seem to be associated with the double deficit type of dyslexia. In future research it seems interesting to examine the relation between EEG data and the phonological or orthographic deficits. Outcomes of these studies might further contribute to the diagnosis of subtypes of dyslexia.
Finally, this study demonstrated that increased EEG power could not explain the increased coherence findings in dyslexia, suggesting these measures reflect different neural networks. The positive correlations between coherence and the different tests demonstrated that these increased coherences might reflect compensatory mechanisms rather then being part of the real core dysfunction in dyslexia, whereas the increased slow activity might be part of the core dysfunction in dyslexia. This should be taken into account in future studies that are trying to elucidate dysfunctional networks in dyslexia. These dysfunctional networks can be dissociated from acquired compensatory mechanisms. Also treatments focused on normalizing brain function (e.g. rTMS, EEG Biofeedback or Neurofeedback) will benefit from this since they could target the deficit rather then target acquired compensatory mechanisms.