home index by author index by subject

Talk;imaging

Saturday stream 2 Session 11.35 - 12.50 Length 25 minutes

An Investigation into the Visual and Auditory Processing Mechanisms Impaired in Dyslexia using Event Related Brain Potentials

Aditi Shankardass, Roderick I Nicolson and Angela J Fawcett

Department of Psychology, University of Sheffield Sheffield S10 2TP, UK. a.shankardass@shef.ac.uk

Abstract

Recent research suggests that dyslexia may result not from a deficit in the brain's ability to process language but to process general visual and auditory information, which may then contribute to the language difficulties. However it is still not certain whether this processing deficit is solely at the sensory stage or also the cognitive stage and in the case of a sensory deficit whether it is the spectral or temporal features of auditory stimuli that cause difficulty. Our aim is to isolate the various stages at which the processing of various aspects of visual and auditory stimuli is impaired in the same set of dyslexic children by studying the brain's event related potentials (ERPs).

In two separate studies, ERP's were recorded in 8 dyslexics and 8 controls during visual and auditory oddball tasks. In the visual study the subjects actively attended to the stimuli and responded to an infrequently (20%) presented target shape (cross) and ignored the frequently (80%) presented standard shape (circle). In the auditory study there were 2 stimulus conditions, where the target tones (20%) differed from the standard tones (80%) in frequency and duration. There were 3 response conditions: passive, where the subjects did not pay attention to the ongoing tones, active, where the subjects actively attended to the tones and responded to the targets and learnt, where the subjects again ignored the ongoing tones. In both studies ERPs were analysed and a dyslexic-control comparison of the following ERP components was made for each stimulus condition: the P1 and mismatch negativity (MMN) components in the auditory passive task (reflecting pre-attentive, sensory discrimination); the N2 and P3 component in the visual task and auditory active task (reflecting attentive, cognitive recognition) and the MMN component in the auditory learnt task (reflecting learning through previous attentive practice).

Results from the visual study and the active condition of the auditory study showed that dyslexics had a significant delay in the peak latency of the P3 component in both the visual shape and the auditory pitch and duration conditions but no delay in the preceding N2 component. This suggests a deficit specifically confined to cognitive processing that is not due to a knock-on effect from any sensory processing deficit and is not specific to modality: dyslexics are slower in the conscious categorisation of visual and auditory stimuli, that is, the conscious recognition and selection of stimuli for the purposes of a response.

I. Background

1. Dyslexia: a general processing deficit?

• recent evidence suggests that dyslexia may result not just from a specific deficit in the processing of language but a more basic deficit in the processing of visual and auditory information
• now substantial evidence of slow information processing by dyslexics on many non-linguistic visual and auditory tasks (Wolf and Bowers, 1999)
• however it is still not clear where the 'bottleneck' in the processing of visual and auditory information is in the brain

2. Possible causes of a general processing deficit in dyslexia

Sensory Processing Deficit

• single bottleneck at the level of the primary sensory cortex
• considerable evidence of sensory deficits with dynamic stimuli: difficulty detecting stimulus timing and change

• with visual stimuli: impaired flicker and coherent motion detection (Talcott et al., 1998)
• with auditory stimuli: impaired temporal (rapid transient) discrimination (Tallal, 1980) and spectral (low level frequency modulation) detection (McAnally and Stein, 1996)

• advocated by proponents of the Magnocellular Deficit Hypothesis (Tallal et al., 1993; Galaburda & Livingstone, 1993; Stein 1997): abnormalities in the pathways that carry rapidly-changing information from the sensory organs to the visual and auditory cortex
• strong relationship between sensory sensitivity and reading performance (Talcott et al., 2000)

• visual motion sensitivity and orthographic skill
• auditory FM sensitivity and phonological skill

Cognitive Processing Deficit

• additional bottleneck in cognitive processing at the level of higher-order association cortex
• difficulty could lie at the attentive stages of recognition, categorisation and response selection
• evidence of reduction in the processing speed for all types of stimuli, including those for which dynamic visual and auditory sensitivity is unlikely to be involved (Denkla and Rudel, 1976; Nicolson and Fawcett, 1994)
• advocated by proponents of the Cerebellar Deficit Hypothesis (Fawcett, Nicolson & Dean, 1996): problem could be related to abnormalities in the cerebellum, which may be responsible for automatisation of tasks

3. Technique employed

High-density event related brain potentials

• electrophysiological brain imaging method to study temporal aspects of processing
• relates EEG recorded and analysed at scalp to neural activity at cortical surface

Wave forms of interest in the EEG trace

4. Research objectives

• investigation of the processing deficit in both visual and auditory modalities
• in the same subjects with dyslexia

• isolation of the stages at which speed of processing of basic non linguistic information is impaired in dyslexics by monitoring ERPs during the task and measuring latency to peak/ amplitude differences for relevant waveforms:

II. Visual Study

Predictions

1. Participants

Table 1: Mean age, IQ and reading age for participants

2. Research Design

Stimulus presentation

• based on the oddball paradigm: visual target stimuli (X) randomly interspersed among standard stimuli (O)
• 200 trials: target tone presented 20% trials and standard tone presented 80% trials
• requires selective choice: responding to infrequent target and ignoring frequent non-target

Data Acquisition

• response time and accuracy recorded

EEG Recording

• brain electrical activity measured using dense array Geodesic Sensor Net comprising of 64 electrodes evenly distributed across the scalp

Fig 1: 64 electrode Geodesic Sensor Net placed on participant's head

3. Data analysis

• segmenting to 100ms before stimulus onset, 900ms after
• digital filtering of trials for artifacts and eye blinks (lowpass 30Hz, highpass 10Hz)
• averaging over all 40 relevant trials
• re-referencing to the linked mastoids
• baseline correction
• bad channel replacement

average latency to peak for P1 waves for relevant trials in visual cortex measured from electrode 38, corresponding to electrode Iz on 10-20 electrode system average latency to peak for P3 waves for relevant trials in parietal cortex measured from electrode 65, corresponding to electrode Cz on 10-20 electrode system calculated for each subject in both groups independent, 1-tailed, t-tests conducted

4. Results

Voltage maps across scalp showing peak P3 activity in controls & dyslexics in early P3 time window

Fig. 5 Voltage maps for dyslexics and controls at 436 msec

P1 latency to peak has no significant difference in dyslexics and controls [t (12) = - 1.4, NS]

P3 latency to peak is significantly longer in dyslexics
mean difference 52 ms [t (12) = 3.4; p<0.01]

III. Auditory Study

1. Participants

The same as those in the Visual Study.

2. Research Design

Stimulus presentation

• pure tones delivered binaurally in pseudorandom order
• based on the oddball paradigm: target stimuli randomly interspersed among standard stimuli
• each of the 4 stimulus blocks consists of 200 trials
• target tone presented 20% trials and standard tone presented 80% trials

Stimulus Conditions

Response Conditions

3. Preliminary Results: Active Response Condition

N2 latency to peak has no significant difference in dyslexics and controls

P3a latency to peak is significantly longer in dyslexics for the far pitch deviant and far duration deviant conditions {PF: [t (10) = 2.27 p<0.05] DF: [t(10) = 1.93 p<0.05]}

P3b latency to peak is significantly longer in dyslexics for the near pitch deviant and near duration deviant conditions {PN: [t (10) = 1.81 p=0.05] DN: [t(10) = 1.74 p=0.05]}

This experiment has just been completed recently
Passive 1 and Passive 2 conditions still need to be analysed
Not yet a complete picture

Interpretation

• the deficit is also not attributable to a knock on effect from earlier sensory delays since the earlier sensory peaks have no significant latency differences
• the deficit is not attributable to motor response selection or execution since it is observed in the earlier time window which reflects classification of the stimuli
• hence the deficit appears to be linked to response categorisation or the need to make a conscious discrimination between stimuli for the purposes of a response
• this is observed in both the visual and auditory domain suggesting possible differences in higher cognitive functions that are not specific to modality
• this would suggest that speed deficits in general non linguistic discrimination tasks in dyslexia lie in the classification and not (just) the perception of the stimuli
• this does not preclude the possibility of sensory deficits in the same set of subjects and in the auditory study, an analysis of the MMN differences to both pitch and duration deviants will highlight any possible impairment in the pre-attentive discrimination of both spectral and temporal aspects of the stimuli