Visual perception and frontal lobe in intellectual disabilities: a study with evoked potentials and neuropsychology
Abstract
Background Perception disorders are frequently observed in persons with intellectual disability (ID) and their influence on cognition has been discussed. The objective of this study is to clarify the mechanisms behind these alterations by analysing the visual event related potentials early component, the N1 wave, which is related to perception alterations in several pathologies. Additionally, the relationship between N1 and neuropsychological visual tests was studied with the aim to understand its functional significance in ID persons.
Method A group of 69 subjects, with etiologically heterogeneous mild ID, performed an odd-ball task of active discrimination of geometric figures. N1a (frontal) and N1b (post-occipital) waves were obtained from the evoked potentials. They also performed several neuropsychological tests.
Results Only component N1a, produced by the target stimulus, showed significant correlations with the visual integration, visual semantic association, visual analogical reasoning tests, Perceptual Reasoning Index (Wechsler Intelligence Scale for Children Fourth Edition) and intelligence quotient.
Conclusions The systematic correlations, produced by the target stimulus in perceptual abilities tasks, with the N1a (frontal) and not with N1b (posterior), suggest that the visual perception process involves frontal participation. These correlations support the idea that the N1a and N1b are not equivalent.
The relationship between frontal functions and early stages of visual perception is revised and discussed, as well as the frontal contribution with the neuropsychological tests used. A possible relationship between the frontal activity dysfunction in ID and perceptive problems is suggested. Perceptive alteration observed in persons with ID could indeed be because of altered sensory areas, but also to a failure in the frontal participation of perceptive processes conceived as elaborations inside reverberant circuits of perception–action.
Introduction
Perception deficits in persons with intellectual disability (ID) have been studied for many years. Ellis (1963) in ‘Handbook of Mental Retardation’ devoted two chapters explicitly to revise how their perceptive problems influence their cognitive difficulties. However, research on perception losses associated to ID was relatively ignored from the mid 1970s for several years. (Fox & Oross 1992).
Several studies found an acute deficit in shape perception capacity (Hermelin & O'Connor 1961a,b), stereopsis vision (O'Connor & Hermelin 1961; Fox & Oross 1988) and the detection of changes in objects colour, shape or even presence and absence (Carlin et al. 2003) in groups of persons with ID.
In a review of their research programme on mild ID perception disorders Fox & Oross (1992) concluded these deficits were because of an early alteration of perception mechanisms that they attributed to neural deterioration and not to optical problems, motivation or understanding.
Stiers et al. (1999) and Stiers et al. (2001) found, in studies in children with early brain damage from different origin, a high correlation between the Frostig Perceptual Quotient and nonverbal intelligence quotient (IQ) [Wechsler Intelligence Scale for Children-Revised (WISC-R)], and they concluded that this pathology interferes with specific visual subsystems function. The authors inferred that the perception organisation problems affected intelligence. In this respect, Courbois (1996) attributed to visual perception problems the limited capacity for mental imagery that would affect higher cognitive functions.
The interpretation of a primary perception alteration is supported by research performed with evoked potentials during the study of immediate brain response to visual stimuli. Osaka & Osaka (1980), Psatta (1981), Hakamada et al. (1981), Gasser et al. (1988) and Zurrón & Díaz (1995) described abnormalities in visual evoked potentials in this population. Furthermore, research with event-related potentials (ERPs), in studies of later and more complex stages of the visual processing have also found several anomalies in these populations (Sandman & Barron 1986; Muñoz-Ruata et al. 2000; Shoji et al. 2002.)
ERPs provide a valid electrophysiological method to study the successive stages of information processing. It was shown repeatedly that ERPs provide reliable indexes to assess cognitive function (Sinha et al. 1992; Trejo et al. 1995; Olvet & Hajcak 2009a,b). However, ERPs have not been used widely to diagnose and understand ID.
Within visual ERPs, both the N1 wave as the one presented immediately before, the P1, have been considered representing the primary visual response. P1 wave was decreased in amplitude and delayed in latency in people with ID with different aetiology and learning disorders in most research studies (Lux 1977; Thun-Hohenstein et al. 1992; Pietz et al. 1996; Brecelj et al. 1996; Brannan et al. 1998; Schulte-Körne et al. 1999).
The pioneering work that, discusses the N1 wave appeared in the 1970s with passive visualisation of scenes (Schulman-Galambos & Galambos 1978). Subsequently, several exploration models have been used: letter recognition (Warren & Wideman 1983), orientation and selection (Harter et al. 1989), exploration of foveal and peripheral vision (Neville & Lawson 1987), shape, colour and movement in both selection and orientation models (Anllo-Vento & Hillyard 1996), passive face recognition (Pineda et al. 1994; Csibra et al. 2008), illusory figures as those of Kanizsa (Herrmann & Bosch 2001; Proverbio & Zani 2002; Grice et al. 2003; Senkowski et al. 2005; Brodeur et al. 2008), visual oddball paradigm (Smit et al. 2007) and comparing animal's images with artificial objects (Proverbio et al. 2007).
N1 refers to two waves that appear at different times and in different locations. The frontal N1 wave was initially considered as the first part of the one that appeared later in the occipito – temporal electrodes. However, soon afterwards researchers began to find functional differences between the anterior and posterior electrodes (Luck et al. 1994; Johannes et al. 1995; Luck & Hillyard 1995). Using an independent component analysis, Makeig et al. (1999) identified two early independent negative waves; one with anterior localisation and another with posterior location. Saron et al. (2001) found that frontal activation occurred with simple visual stimuli. By analogy with studies of macaques this group attributed the formation of the frontal N1 to the arrival of information by visual pathways that, according to Foxe & Simpson (2002), reach the prefrontal by the dorsal pathway about 10 ms before that by the ventral pathway.
The frontal N1 wave reverses its polarity in multiple sclerosis, and this is interpreted as a consequence of demyelination of the fronto-occipital neurons (Gonzalez-Rosa et al. 2006). In spite of the clear distinction between the two wave components, many studies have only addressed the N1b wave (Heinze et al. 1990; Wijers et al. 1993, 1997; Talsma & Kok 2002; Wang & Kameda 2005; Fu et al. 2008).
The level of correlation between visual ERPs and intelligence tests has been variable. Barry & Ertl (1966) found negative correlations, ranging from −0.88 to −0.76, between latencies of these potentials and intellectual level. However, Shucard & Horn (1972) found coefficients between −0.15 and −0.32 between latencies and several psychological tests.
Specifically in the waves we are interested in, Sandman & Barron (1986) compared two groups of people with diminished intelligence, finding that those with higher intelligence level had significantly shorter latencies in visual P1 and N1 than those with lower intelligence level. With regard to the amplitude, Haier et al. (1983) obtained significant correlations between the N1–P2 amplitude and intelligence level. On the other hand, Patterson et al. (1989) found no significant correlation between the N1 amplitude and intelligence psychometric tests, but the difference between the attended and unattended N1 amplitude produced significant correlations with scores on perceptual-motor tests and figures memory.
Relationships between the N1 wave and cognitive impairment have been studied by Reza et al. (2003) in 30 patients with posttraumatic brain injury. They obtained correlations between the N1a wave latency and the Mini-Mental State Examination (MMSE) (−0.45) and the executive IQ (−0.56).
Objectives
The aim of this work is to determine if the N1 wave can be obtained in a sample of people with ID and, if so, to understand the relationship with neuropsychological tests. These data could clarify if alterations in perceptual mechanisms could lead or contribute to cognitive deficits.
If the N1 wave is related to complex processes, beyond a mere entry of visual stimuli, such a wave should correlate with measures of perceptual organisation as well as with intellectual level tests and other forms of perceptual reasoning. This in turn would indicate that at least some people with ID may have alterations of the perceptual organisation that would show up in the N1 wave.
Methods
Participants
We studied 77 children enrolled in a Foundation with the goal to promote education and employment of people with ID, and obtained prior informed consent from their parents or guardians. Eight subjects were excluded, three because of uncontrolled artefacts and five did not complete the task, leaving the studied sample in 69 subjects.
The most common aetiologies of our subjects ID, as reflected in hospital reports and following the Diagnostic and Statistical Manual of Mental Disorders Fourth Edition Text Revision (American Psychiatric Association 2000) classification were: pregnancy and perinatal pathologies (43.47%), genetic diseases (26.08%) medical conditions acquired in childhood (4.34%), environmental causes (17.39%) and unknown causes (8.69%).
As suggested by Gasser et al. (2003) we expected the aetiological heterogeneity of the sample would increase the data variability, both neurophysiological and psychological, improving the contrasts between the two types of measures. The ages of the subjects were between 12 and 17 years old with a mean age of 14.98 years (SD 1.7). Thirty-nine subjects were male and 30 female.
Twenty subjects (28.9%) took medication (antiepileptic six, neuroleptics six, antidepressants one, stimulants one and combinations of the above six). Participants did not stop taking their medication before the examination. All subjects with refractive problems used their corrective glasses. Most were right-handed, while 8% were left-handed.
Neurophysiological evaluation
Stimulus
The paradigm of discrimination and selection of shapes with odd-ball presentation was used to obtain the N1 as by Almasy et al. (1999), Vogel & Luck (2000) and Smit et al. (2007).
The stim 2 software (Compumedics Ltd. El Paso, TX, USA) was used to present four geometric shapes: a diamond, a circle, a triangle and a 5-pointed star all in black over a white background. The target stimulus was the triangle. A series of 250 stimuli, out of which 20% matched the target, were presented randomly. The stimulus duration was 700 ms with an inter-stimulus interval of 1500 ms.
Data acquisition
Subjects were instructed to respond to the target stimulus by pressing the Stimpad (Compumedics). The test was carried out in an isolated room. The subjects were monitored to keep them alert, calm and looking at the screen.
The acquisition and analysis of records was carried out with a Neuroscan amplifier ‘SynAmps2’ and the software Scan 4.4 (Compumedics).
A 32-channel standard electrode cap was used with chlorinated silver electrodes (AgCl). The distances between electrodes were constant and according to the IFCN (International Federation of Clinical Neurophysiology) (Klem et al. 1999) following the 10-20 international system. The electrodes positions were FP1, FP2, F7, F3, Fz, F4, F8, FT7, FC3, FC4, FT8, T3, C3, Cz, C4, T4, TP7, CP3, CP4, TP8, T5, P3, PZ, P4, T6, O1, O2 and Oz. The reference was placed on the forehead. Two bipolar derivations were used to monitor vertical and horizontal ocular movements through its electrooculogram.
The electroencephalogram for each electrode was digitised at 1000 Hz and low and high pass filters were 100 Hz and 0.30 Hz, respectively. The artefact rejection criterion was any positive or negative change above 75 µV. The length of each record (epoch) was of −100 ms to 1000 ms. The baseline was corrected from −100 ms to the 0 ms. The electrode impedances were kept constant and below 5 kΩ for all electrodes and below 10 in the electrooculogram.
Analysis
To obtain the ERPs, the artefact-free epochs were used. In this way we obtained two averages, one for the target stimulus and one for the non-target.
To define the waves, we used both the latency and amplitude, as well as the polarity and topography following the view of Donchin et al. (1978) according to which the polarity and topography of the ERPs indicate the source of the wave. The amplitude and latency are related to psychological functions.
We define the N1a peak, according to the time window criteria of Smit et al. (2007), as the largest negativity following the P100 wave, with latency between 88 and 168 ms located in the frontocentrals and parietal electrodes. N1b peak was the largest negativity, between 132 and 220 ms, located in the occipito-temporal electrodes. For both peak-to-baseline amplitudes were scored. The peaks of the waves were checked visually to make sure they corresponded to the desired wave.
Neuropsychological assessment
The intellectual level of the subjects was determined using the Wechsler Intelligence Scale for Children Fourth Edition, WISC-IV (Spanish version, Wechsler 2005). This test provides a Full Scale Intelligence Quotient IQ (FSIQ) and four indexes: verbal comprehension index, perceptual reasoning index, working memory index and processing speed index. In this version of the WISC the traditional verbal and manipulative quotients have disappeared.
To evaluate the selected visual perceptive abilities in detail, the Bender (1981) gestalt visual-motor test corrected by the Koppitz (1994) system was used. The correction considers the following errors in the models copy: distortions, rotations, perseverations and integration of the shapes. The Motor-Free Test of Visual Perception (MVPT-3) Colarusso & Hammill (1980), which is a test that has proven sensitive to visual agnosias (Lundberg 2003), was used.
For the visual reasoning evaluation, the following subtests were selected from the Illinois Test of Psycholinguistics Abilities, ITPA (Spanish version, Kirk 2004): Visual Reception, which measures the ability to get the meaning of visual images, choosing from a set of four pictures the one conceptually similar to the stimulus image; and Visual Association which measures the ability to relate concepts presented visually by inferring the rule that relates them.
These tests can be used in persons with ID, with similar levels of intelligence to those of the studied sample.
Statistical methods
The results of the neuropsychological tests from 69 valid cases were collected and analysed statistically. The reaction times and errors committed with the geometric stimulus and voltages and latencies of the N1a and N1b waves produced both by the target and non-target stimulus, were also recorded.
The adjustment to a normal curve of the data was tested and the means and standard deviations (SD) were calculated.
Later, the linear correlations between neuropsychological and neurophysiological variables were calculated with the aim to provide ideas on the possible functional significance of the waves, as suggested by Altenmüller et al. (2004). These correlations were made with the N1a and N1b waves for both target and non-target stimuli.
Results
Descriptive statistics
Psychometric data
The 69 subjects with valid data had a FSIQ mean of 55.9 (SD 10.26). The means and deviations of the indexes subdividing the WISC-IV test range from 61.6 (SD 9.6), for the verbal comprehension index, to 64.5 (SD 14.7) for the working memory index. The average level of the subjects falls therefore within the mild ID.
The average developmental age according to the Bender Test was 7.9 years (SD 1.5) being the most frequent errors the distortion and the integration of the figures.
The average developmental age, according to the Colarusso-Hammill Motor-Free Test of Visual Perception (MVPT-3), was of 8.2 years (SD 1.3) with age fluctuations in the subtests that composed the test between 7 and 8 years and a half.
In the subtests selected from the ITPA, Visual Reception and Visual Association, subjects reached a developmental mean age of 8.7 years and of 9.4 years, respectively.
Neurophysiological data
The mean peak amplitudes and latencies of the N1a (Fig. 1) are very similar for both target (Fz − 3.6 µV, SD 5.3, 115.8 ms, SD 21.3) and for non-target (Fz − 3.9 µV, SD 3, 116.3 ms, SD 22) stimuli.
Average event-related potentials waveforms elicited by targets (
) and non-targets (
) in the Fz electrode.
The N1b produced by the target stimulus (Fig. 2) had a slightly higher mean peak amplitude and a somewhat lower mean latency (Oz − 5.9 µV, SD 9.1, 182 ms, SD 29.7) than those produced by the non-target stimulus (Oz − 4.9 µV, SD 5.6, 187 ms SD 25.5). The differences between these averages were not statistically significant.
Average event-related potentials waveforms elicited by targets (
) and non-targets (
) in the Oz electrode.
Significant gender differences were not found in any case for peak amplitudes or latencies.
The mean of N1b peak amplitudes, in our subjects, were similar to those obtained by Smit et al. (2007) using healthy young subjects. However, both N1a and those of N1b latencies in our sample have a statistically significant delay compared with those of Smit et al. (d.f. = 398, t = 4.35, P < 0.01 and t = −6.64, P < 0.01) The peak amplitudes of our N1a waves were also significantly wider than those of Smit (t = 6.89, P < 0.01).
Medication
Mean differences were performed, in all electrodes, between subjects taking neuroleptics only, anticonvulsants only, combined therapies and all previous groups together without finding significant differences compared with subjects with no medication.
Average response time and errors
The average response time to the target stimulus was 545 ms (SD 96) with a range between 349 and 717 ms. The average for omission of response errors was 4.6% in all trials while in 3.9% the subject pressed the button by mistake (commission error).
Correlation statistics
No significant correlations with neuropsychological test scores were found in the peak amplitudes or in the latencies of N1b waves, both target and non-target, or in the non-target N1a waves. In all cases a sporadic and non-systematic low level significant correlation may have appeared by chance.
There were also no significant correlations between any of these waves and the age of the subjects. Since Smit et al. (2007) found differences between old and young participants, perhaps the constrained age range of this study means that the range was not available for the correlations to be evident.
However, systematic significant correlations were obtained at electrodes around Fz related to the N1a wave. A correlation was found between N1a peak amplitude produced by target stimuli and several neuropsychological test scores (Table 1), but not with latency.
| WISC_FSIQ † | WISC_PR ‡ | Block design | Picture concepts | Matrix reasoning | Picture completion | |
|---|---|---|---|---|---|---|
| C4 amplitude N1 | 0.207 | 0.266* | 0.267* | 0.215 | 0.211 | 0.317* |
| 0.109 | 0.038 | 0.038 | 0.096 | 0.103 | 0.022 | |
| Cz amplitude N1 | 0.250* | 0.301* | 0.270* | 0.214 | 0.208 | 0.321* |
| 0.050 | 0.017 | 0.034 | 0.095 | 0.104 | 0.020 | |
| FC4 amplitude N1 | 0.301* | 0.373** | 0.311* | 0.333** | 0.335** | 0.354* |
| 0.017 | 0.003 | 0.014 | 0.008 | 0.008 | 0.010 | |
| C3 amplitude N1 | 0.212 | 0.259* | 0.268* | 0.223 | 0.189 | 0.229 |
| 0.099 | 0.042 | 0.035 | 0.081 | 0.141 | 0.103 | |
| Fz amplitude N1 | 0.332** | 0.404** | 0.351** | 0.352** | 0.289* | 0.305* |
| 0.008 | 0.001 | 0.005 | 0.005 | 0.023 | 0.028 | |
| F4 amplitude N1 | 0.336** | 0.400** | 0.314* | 0.365** | 0.280* | 0.295* |
| 0.008 | 0.001 | 0.013 | 0.004 | 0.028 | 0.034 | |
| FC3 amplitude N1 | 0.238 | 0.295* | 0.273* | 0.300* | 0.228 | 0.312* |
| 0.063 | 0.020 | 0.032 | 0.018 | 0.075 | 0.024 | |
| F3 amplitude N1 | 0.273* | 0.315* | 0.248 | 0.351** | 0.218 | 0.206 |
| 0.033 | 0.013 | 0.054 | 0.005 | 0.091 | 0.147 |
- * P < 0.5,
- ** P < 0.01 (two-tailed test).
- † Wechsler IV Full Scale Intelligence Quotient.
- ‡ Wechsler Perceptual Reasoning Index.
The possible error derived from multiple comparisons tends to be compensated because on the one hand, correlations point to brain zones identified previously by other methodologies and, on the other hand, the probability that correlations only appear with visual tests and only in the electrodes of the relevant areas is rather small.
WISC-IV
With regard to the WISC-IV, maximum correlations were obtained between the Perceptual Reasoning Index and the peak amplitudes of the electrodes around where the N1a wave was produced with a maximum coefficient in Fz of 0.404. There were no significant correlations with other WISC-4 indexes. Obviously, there were correlations with the FSIQ, of which the Perceptual Reasoning Index is a part.
The perceptual reasoning subtests presented systematic significant correlations of the N1a peak amplitude with a higher intensity in the Block Design and Picture Concepts than in the Matrix Reasoning and Picture Completion subtests. These correlations were positive in all cases (Table 1).
Bender
In the Bender test, correlations in the N1a peak amplitude occurred mainly with the number of integration of form errors, but not with the rotation, distortion or perseveration errors. The correlations, which appeared centred at electrode Fz, were negative in this case, because unlike in the other tests, the score is the errors number (Table 2). The integration errors committed by the subjects typically consisted of drawing figures breaking the gestalt (simultanagnosia).
| Test | Colarusso – Hammill † | Bender ‡ | ITPA § | |
|---|---|---|---|---|
| Subtest | Visual closure | Integration | Visual reception | Visual association |
| C4 amplitude N1 | 0.398** | −0.212 | 0.228 | 0.459** |
| 0.002 | 0.107 | 0.083 | 0.000 | |
| Cz amplitude N1 | 0.448** | −0.239 | 0.284* | 0.486** |
| 0.000 | 0.066 | 0.028 | 0.000 | |
| FC4 amplitude N1 | 0.384** | −0.226 | 0.261* | 0.489** |
| 0.003 | 0.083 | 0.044 | 0.000 | |
| C3 amplitude N1 | 0.421** | −0.255* | 0.315* | 0.452** |
| 0.001 | 0.050 | 0.014 | 0.000 | |
| Fz amplitude N1 | 0.265* | −0.279* | 0.316* | 0.503** |
| 0.045 | 0.031 | 0.014 | 0.000 | |
| F4 amplitude N1 | 0.298* | −0.263* | 0.285* | 0.505** |
| 0.023 | 0.041 | 0.027 | 0.000 | |
| FC3 amplitude N1 | 0.306* | −0.264* | 0.358** | 0.500** |
| 0.019 | 0.042 | 0.005 | 0.000 | |
| F3 amplitude N1 | 0.190 | −0.193 | 0.336** | 0.434** |
| 0.157 | 0.143 | 0.009 | 0.001 | |
- * P < 0.50.
- ** P < 0.01 (two-tailed test).
- † Colarusso – Hammill, the Motor-Free Test of Visual Perception (MVPT-3).
- ‡ Bender gestalt visual-motor test.
- § ITPA, Illinois Test of Psycholinguistic Abilities.
MVPT
The MVPT-3 of Colarusso and Hammill presented significant correlations with the N1a in the Visual Closure subtest with a maximum at the Cz electrode. This test scored the number of incomplete schematic images that the subject was able to recognise (Table 2). The Visual Discrimination, Visual Memory and Spatial Relationship subtests showed no correlations reaching statistical significance.
ITPA
Within the ITPA, high correlations were found systematically with the Visual Association Test in all electrodes that record the N1a wave. The Visual Reception Test also showed systematic correlations with these same electrodes with a slightly lower significance level (Table 2).
Neuropsychological tests significantly related to N1a, were also correlated with coefficients ranging between 0.54 and 0.40 and being significant in all cases.
These results match those obtained by Reza et al. (2003) in adult patients with brain damage in which neither the latencies nor the N1b wave amplitudes correlated with neuropsychological tests and yet the latency of the N1a, and not the amplitude as we observed, correlated with both the MMSE and with the executive IQ (Wechsler Adult Intelligence Scale – Revised), which shares with the perceptual reasoning index of the WISC-IV the Block Design and Picture Completion subtests.
Discussion
Our results suggest frontal involvement in both the integration of visually perceived forms as well as in images cognition. Frontal involvement in visual perception processes has been studied from different points of view. Under an anatomical point of view, Cavada and Goldman-Rakic (1989), Petrides & Pandya (1999) and more recently Cavada et al. (2000) among others, described anatomical pathways connecting the different sensory visual areas with the prefrontal areas in macaques. In primates, these connections reach the prefrontal cortex through the dorsal and ventral visual pathways, starting from multiple areas of the extra-striate cortex (Schall 1997) and from the thalamic nuclei (Asanuma et al. 1985; Garey et al. 1991). In humans, the communication of visual areas with the prefrontal cortex through the dorsal and ventral visual pathways was described, by, among others, Foxe & Simpson (2002) with electrophysiological methods and by Gregoriou et al. (2009), who found coupling between the visual and prefrontal cortex in the gamma frequencies.
A question that might arise is whether the N1a wave reflects the impulse of the motor response or only visual perception related phenomena. Saron et al. (2001) used dipole studies and intracranial electrodes that showed the independence of these potentials and the task motor response. The results of our study support this idea, as we did not see an N1a wave lateralisation to the contra-lateral side of the hand which executes the response; the wave appeared, on average, 430 ms before the motor response and also with the non-target stimulus to which there is no motor response.
Frontal involvement in visual perception has been interpreted in different ways. It is often assumed that the role of the frontal lobe in the visual processing is to control attention (Johannes et al. 1995; Vogel & Luck 2000; Marois et al. 2004; Kehrer et al. 2009). In this sense, some studies found that the N1 amplitude increases in attention deficit disorders (Prox et al. 2007), as well as in reading disability (Harter et al. 1989). Strandburg et al. (1984) found that the N1 amplitude increased with the difficulty of the task in his control group.
In our case, taking into account the sign of the correlations, wider amplitude (negative sign) in the N1a wave implies a lower IQ. Because most of the subjects solved correctly the task used as a stimulus, it could be hypothesised that the increase in N1 is because subjects with lower IQ need more effort (Strandburg et al. 1984; Mulert et al. 2005).
A slightly different view is supported by Barcelo et al. (2000) who observed that prefrontal damage reduced neuronal activity in the extrastriate visual cortex. According to these authors, when the prefrontal controls the visual attention, perceptive templates are activated, so the selection of the objects relevant features and their placement in a spatial configuration is guided ‘top-down’ to allow their recognition. Therefore, frontal dysfunctions are suggested to disrupt the perception. Summerfield et al. (2006) had a similar position.
The correlations of the N1a amplitude, produced by the target stimulus, with the errors of shape integration in the Bender, and visual closure in the Hammill-Colaruso, tests could be because of alterations in the functioning of the perceptual template system in subjects with ID. If this were the case, the N1 amplitude would probably reflect a momentum of this process that would be interactive and recursive between occipital and frontal areas (Doniger et al. 2001; Foxe & Simpson 2002; Sehatpour et al. 2006).
A difficult issue is why the correlations reached statistical significance only with the target stimuli. One possible interpretation is that responding to target stimuli involves an active vision that would promote the arrival of information to frontal regions more effectively than simply passively viewing the stimuli (Goodale & Milner 1992; Boussaoud et al. 1995; Foxe & Simpson 2002).
Three levels of organisation
Integration and visual closure
The perception of images needed for visual closure (MVPT-3) and for the shape integration (Bender) tests is based in the gestalt grouping of the traits that constitute them. In turn, the block design is considered as a test for analysis and integration of non-semantic visual information (Strauss et al. 2006) in which the performance is mainly determined by recognition of the gestalts models (Lezak et al 2004). In addition, the block design performance correlates with frontal damage (Witgert et al. 2010) in patients with amyotrophic lateral sclerosis. The picture completion test is also a test for integration of perceptual gestalts but unlike the previous, using meaningful images (Lezak et al. 2004).
The study of Seymour et al. (2008) reveals that the integration of local features in a global gestalt involves the inferoparietal cortex and temporal medium gyrus as well as the prefrontal cortex. Therefore, damage in any of these regions causes simultanagnosia, a deficit in perceiving multiples traits or objects simultaneously (Farah 2004).
We have previously appointed the neurophysiological dynamic of visual closure and perceptual integration of gestalts using the occipito-frontal interactive and recursive process (Doniger et al. 2001; Foxe & Simpson 2002; Sehatpour et al. 2006). After the visual closure, the information is again projected to the prefrontal cortex to retrieve the meaning of the pictures from the semantic memory (Wagner et al. 2001).
Semantic relationship
The connections of the semantic relationship tests, Visual Reception (ITPA) and concepts test (WISC-IV), with the prefrontal cortex has been described in several studies. For example, Wagner et al. (2001) found that left inferior prefrontal cortex (Brodman Area 47) mediates in the semantic knowledge control. The activation of this area increases both the difficulty of semantic recovery as well as the level of control required for the task. The involvement of the ventral visual pathway in this process was proposed by Jeannerod & Jacob (2005) to relate the ventral pathway with a semantic way to elaborate the information.
Also, the results of Bunge et al. (2003) support the hypothesis that the ventrolateral prefrontal cortex (Brodmann's area 45–47) interacts with the temporal and parietal cortex to retrieve information that is semantically associated. Their activation is greater the weaker the semantic relationships between stimuli are (Bunge et al. 2004). In addition, Bunge et al. (2005) found activation of the rostrolateral prefrontal cortex (Brodmann's area 10) in tests judging the semantic analogy between visual images.
Analogical reasoning
Complex analogical reasoning, as it is done in the Visual Association test (ITPA) and Matrix Reasoning (WISC-IV) has also been associated with prefrontal cortex areas. In neuroimaging studies, the performance of tasks with analogical reasoning activated rostrolateral prefrontal regions (Wright et al. 2008). Analogous results were found by Crone et al. (2009) using Raven's Progressive Matrices test. Christoff et al. (2001) also found right dorsolateral activation (Brodmann's area 9/46).
To determine to what extent working memory influenced this prefrontal activation, Green et al. (2006) compared brain activation produced by analogical reasoning tests with activation produced by working memory tasks. Analogical reasoning activated the rostrolateral prefrontal cortex, while working memory tasks that, did not involve analogical reasoning activated parietofrontal regions.
The correlations obtained from the N1a wave with closure, association and visual reasoning tests suggests that these processes would be reflected very early in the frontal N1a wave, with an average of 115.8 ms in our sample, and could be the first reverberating circuit momentum of the working memory in perception–action cycle as it is described by Fuster (2008).
Conclusions
The stimuli used in this study appeared suitable to evaluate the subjects: generally children responded within the time provided, and the low number of errors in the tests indicated that subjects remained sufficiently attentive and were able to correctly discriminate geometric figures.
We found that the N1 wave (N1a and N1b) can be obtained with an active discrimination and selection paradigm and that in children with ID, its morphology and topography were similar to the waves described in the literature, although we found differences in amplitudes and latencies.
The finding of systematic significant correlations in neuropsychological tests that assess perceptual abilities with the N1a (frontal) produced by the target stimulus, and not with N1b (posterior), suggests that the visual perception process involves frontal participation. In addition, these correlations support the idea that the N1a and N1b are not equivalent, which is consistent with the findings of Makeig et al. (1999) and Smit et al. (2007). Neither wave produced by the target stimulus can be considered functionally equivalent to those produced by the non-target stimulus, as the correlations are found only with the potentials generated by the first.
Because correlations were only found with tests involving visual perception integration processes: visual closure and integration errors in the Bender test, it could suggest that some of the persons with ID studied have primary perception organisation problems of frontal origin with different intensities causing the correlations.
Moreover, the relationship of the N1a wave with the comprehension and visual association capacities could mean that this wave may reflect deterioration of these complex intellectual functions. Both these correlations, as the one with the perceptual reasoning index from the Wechsler scale, might be the result of impaired visual perception organisation or directly reflect a perceptual reasoning disorder of higher level. The performance in both types of tests would depend on the proper functioning of the frontal cortex in the subjects studied.
Subjects with ID have variable scores in perceptual and visual reasoning tests, which probably depend on the degree of brain function impairment and correlate with the N1a wave. On the other hand, it is known that people with ID, on average, have an abnormal frontal activity shown by increased power in the frontal slow waves irrespective of the disability aetiology (Martín-Loeches et al. 2001; Gasser et al. 2003) and by outburst of frontal delta waves (Watemberg et al. 2003). Thatcher et al. (2005) found that lower intelligence is associated with a lower frontal synchronisation as measured by coherence, with the posterior cortex. This leads us to hypothesise that at least part of the perceptual alterations observed in people with ID may not directly depend on changes in the sensory areas but from a failure of frontal involvement in the process of percepts construction. This again emphasises the importance to exercise executive functions in people with ID.
The N1a wave amplitude may be related to both frontal functional capacity and, according to the results of this work, to integration capabilities, association and visual reasoning. Thus the study of this wave might be useful as support to the diagnosis of these functions. Although these findings represent only one aspect of the perceptual mechanisms that may be altered in people with ID, the fact that early perception alterations are related to the reasoning should be the subject of future studies.
