Printer Friendly

Spatial working memory and intelligence biological correlates.

Abstract

Previous research utilizing brain imaging techniques has indicated that the prefrontal, parietal and occipital areas all play a role in spatial working memory. Previous psychometric research has indicated a positive relationship between performance on working memory tasks and intelligence. The aim of the present study was to integrate these two lines of research by studying the electrophysiological processes underlying spatial working memory across subjects varying in psychometric intelligence. Steady-state probe topography (SSPT) was used to investigate the cortical activity of 12 average and 12 high IQ volunteers during a spatial working memory task. Holding spatial information in working memory resulted in an increase in the steady-state visual evoked potential (SSVEP) latency in frontal areas, while in parietal and occipital areas there was a decrease in SSVEP latency and an increase in SSVEP amplitude. Increasing the memory load of the working memory task was found to decrease SSVEP latency in the occipital, parietal and right temporal areas and decrease the SSVEP amplitude in the right occipital area. The magnitude of these SSVEP amplitude and latency differences were greater for the high IQ group than for the average IQ group, particularly in posterior areas. These results suggest that the areas of the brain involved in working memory are influenced by individual differences in intelligence.

Keywords: Working memory; Intelligence; SSVEP; SSPT; IQ

1. Introduction

Baddeley (1988, 1992a, 1992b, 1995) has defined working memory as a limited-capacity attentional system that provides temporary storage for information that is to be manipulated or retained for a short period of time. This limited capacity attentional system is known as the `central executive' and is supported by at least two slave systems where information is thought to be actively rehearsed. These two slave systems are known as the articulatory or phonological loop and the visuospatial sketchpad (Baddeley, 1988, 1992b; Baddeley & Hitch, 1974).

Researchers have used metabolic and electrophysiological brain imaging techniques in an attempt to identify the areas of the brain involved in working memory. Jonides et al. (1993) employed positron emission tomography (PET) to investigate the regional cerebral blood flow during a spatial working memory task that involved remembering the location of three dots for a few seconds. They reported increased activation in the prefrontal, premotor, posterior parietal and occipital cortices over both hemispheres, but this increased activation was only significant in the right hemisphere. Other PET and functional magnetic resonance imaging (fMRI) studies have used different spatial working memory tasks, but have also reported a similar pattern of cortical activation (Gold, Berman, Randolph, Goldberg, & Weinberger, 1996; McCarthy et al., 1994; Petrides, Alivisatos, Evans, & Meyer, 1993; Smith et al., 1995).

Studies using event-related potentials (ERPs) have also reported that visuospatial working memory tasks activate the parietal and prefrontal cortex in humans. Ruchkin, Johnson, Grafman, Canoune and Ritter (1992) reported that during a spatial working memory task, the most prominent ERP response was a high amplitude, negative slow wave that was largest over the right parietal cortex. This negative slow wave was found to be directly related to the memory load of the task. The effect of memory load became more pronounced with time over the right parieto-temporal scalp, although the negativity was found to become less lateralized to the right hemisphere while the subject was holding the spatial information in working memory. Ruchkin et al. also reported that during both the spatial and phonological tasks there was a frontal negativity that was lateralized to the left hemisphere. It was suggested that this activity was related to the central executive component of working memory. Ruchkin, Canoune, Johnson, and Ritter (1995) and Ruchkin, Johnson, Grafman, Canoune, and Ritter (1997) replicated these findings in later studies.

A study by Silberstein (1997) also provided support for the importance of parietal and prefrontal processes during a spatial working memory task using a technique known as steady-state probe topography (SSPT). SSPT is a brain imaging technique that involves using a probe ERP to investigate changes in the steady-state visually evoked potential (SSVEP), which have been shown to be related to cognition (Silberstein et al., 1990). SSVEPs are cortical responses to rapidly repetitive stimuli that consist of individual sinusoidal components at both the stimulus frequency and multiples of it, whose amplitude and phase should remain relatively consistent over time (Regan, 1989). The rationale for the probe ERP technique involves both the probe stimulus and the cognitive processes wing for the cortical resources that are available (Papanicolaou & Johnstone, 1984).

Silberstein (1995) has proposed that changes in the amplitude and latency of the SSVEP may occur as a result of neuromodulatory differences in cortico-cortical re-entrant loops. These cortico-cortical re-entrant loops have been proposed to play a vital role in regulating the frequency of electroencephalography (EEG) rhythms in the alpha bandwidth (Silberstein, 1995). The connections involve a cortico-cortical feed-forward (predominantly excitatory) and a cortico-cortical feedback (predominantly inhibitory) connections that are relayed via layer 1 of the neocortex (Silberstein, 1995).

SSPT has been proposed to have several advantages over other brain imaging techniques such as PET and ERPs. One of these advantages is that only between 1 and 5 s of a recorded signal is required in order to estimate the SSVEP amplitude and phase (Silberstein, Ciorciari, & Pipingas, 1995). SSPT also enables the cortical activity to be continuously recorded from a few seconds to several minutes (Silberstein et al., 1990). Therefore, SSPT may be a useful technique for investigating the rapid changes in brain electrical activity while undertaking cognitive tasks (Silberstein et al., 1995). The second advantage of SSPT is that the analysis of the electrical signal is based only on the frequency of the probe stimulus. Therefore, it is less susceptible to interference by noise and artifacts than other techniques (Regan, 1989).

Silberstein (1997) has used SSPT to investigate brain activity during a spatial working memory task compared to a perceptual control task, and reported that during the retention interval of the spatial working memory task there was an increase in SSVEP amplitude in the prefrontal and parietal areas and a decrease in the SSVEP latency in the prefrontal area.

Other studies have also reported an increase in alpha amplitude during tasks that require actively holding and manipulating information. For example, Ray and Cole (1985) reported significantly more alpha power during `rejection' tasks that required mental imagery than during `intake' tasks that involved attending to external stimuli. Similar results were reported during an auditory memory task (Krause, Lang, Laine, Kuusisto, & Porn, 1996) and a spatial working memory task (Okada & Salenius, 1998). Krause et al. reported that presenting the memory set elicited an event-related synchronization, or an event-related enhancement of amplitude (Pfurtscheller, 1992), in the 8-10 and 10-12 Hz EEG bands that continued until the probe was presented. Okada and Salenius (1998) employed magnetoencephalography to investigate neural activity during a spatial working memory task. They reported a suppression of alpha activity (approximately 6-10 Hz) and an enhancement of mu activity (approximately 10-14 Hz) during the retention interval of the task. In contrast, Gevins, Smith, McEvoy, and Yu (1997) reported a decrease in both low and high EEG alpha power during spatial and verbal working memory tasks.

The reported decrease in SSVEP latency during the working memory task reported by Silberstein (1997) was partially consistent with the findings reported by Wilson and O'Donnell (1986) who correlated the rate of memory scanning and the mental rotation of an image with the apparent latency of the SSVEP. They reported that subjects who scanned a memory list and who were able to quickly mentally rotate objects recorded shorter latencies elicited by a 40-59 Hz flickering stimulus. They also reported no significant correlations between performance and SSVEP latency elicited by a stimulus flickering at 15-23 Hz.

In general, research investigating the cortical areas activated during spatial working memory tasks have found that activation tends to be greater in the right hemisphere and the major areas of activation to include the prefrontal cortex, premotor area, posterior parietal area and the occipital lobe. However, these brain imaging studies have tended to use highly educated samples, which may have limited generalizability, as there is substantial evidence for a relationship between working memory and intelligence.

Pellegrino and Glaser (1979) reported that performance on an inductive reasoning task was influenced by an ability to understand relationships between objects, and to be able to hold the outcomes of each step of analysis in working memory. Carpenter, Just, and Shell (1990) proposed that the Raven Advanced Progressive Matrices (APM) measures both the ability to break down items so that they are easier to manage and also higher level abstract reasoning ability. Because the difficult problems on the APM possess a greater number of rules, those individuals who retain more rules and goals in their working memory were better able to solve more complex problems (Carpenter et al., 1990).

Correlational research investigating the relationship between working memory capacity and reasoning ability has indicated that the percentage of variance in reasoning ability explained by working memory capacity ranges from approximately 20% to 80% (Embretson, 1995; Kyllonen & Christal, 1990; Necka, 1992). Similar results have been obtained in research investigating the relationship between working memory capacity and comprehension, with correlations ranging from .3 to .5 (Baddeley, Logie, Nimmo-Smith, & Brereton, 1985; Engle, Cantor, & Carullo, 1992; Swanson & Berninger, 1995). Just and Carpenter (1992) concluded that a larger working memory capacity for language is an advantage in comprehension and may lead to a better understanding of the meanings of words and to an improved vocabulary.

A recent study by Smith, Rush, and Gevins (1998) recorded EEG during a working memory task and correlated the various EEG parameters with scores on the Weschler Adult Intelligence Scale-Revised (WAIS-R). Evidence for a relationship between working memory and intelligence was supported by the finding that approximately 64% of the variance in intelligence scores could be accounted for by both the behavioral and neuroelectric working memory variables (Smith et al., 1998). It was also reported that intelligence was found to be positively related to "... the amplitude and reactivity of an EEG spectral feature in the theta band from frontomesial cortex" and the lower alpha band (Smith et al., 1998, pp. 1-40). Smith et al. also reported a negative correlation between WAIS-R intelligence and the peak latency of the P3 evoked potential (EP) that was related to categorizing the stimulus, and a positive correlation between P3 amplitude and WAIS-R intelligence.

Researchers have also proposed that some of the cortical areas that appear to play a vital role in working memory also play an important role in our ability to learn and perform higher cognitive functions (Wickelgren, 1997) which are often regarded as part of our intellect.

Prabhakaran, Smith, Desmond, Glover, and Gabrieli (1997), using fMRI, investigated cortical activity during performance on the Standard and Advanced Progressive Matrices (SPM and APM). Prabhakaran et al. reported an increase in cortical activation during fluid reasoning bilaterally in the dorsolateral prefrontal, premotor, posterior parietal, occipital, inferotemporal and anterior cingulate areas, with the posterior parietal activation lateralized to the left. This finding is consistent with the results of earlier PET studies by Haier and colleagues (Haier, Siegel, Tang, Abel, & Buchsbaum, 1992; Haier et al., 1988), indicating increased activation in the parieto-temporal, occipital and frontal areas during APM performance.

Prabhakaran et al. (1997) also reported that during the analytic reasoning task there was a bilateral increase in cortical activation in the dorsolateral prefrontal and premotor areas, and a predominantly left hemispheric increase in cortical activation in the posterior parietal, inferotemporal and occipital areas. In contrast, they reported a predominantly right hemisphere increase in cortical activation in the dorsolateral prefrontal, premotor, posterior parietal, occipital, inferotemporal and anterior cingulate areas during figural reasoning tasks. Prabhakaran et al. reported that items from the APM activated areas of the cortex also involved in the executive, rehearsal and storage functions of working memory. They proposed that the findings indicated that spatial, object and verbal working memory systems control fluid reasoning ability.

The evidence for a relationship between working memory and intelligence raises the question of whether the areas of cortical activation during a working memory task are mediated by individual differences in measured intelligence. Therefore, the aim of the present study was to further integrate research examining the electrophysiological processes underlying both working memory and intelligence. The SSPT technique was used to investigate the brain electrical activity of subjects of differing levels of intelligence during a spatial working memory task, (similar to that used by Jonides et al., 1993). Based on the results of previous psychological, electrophysiological and imaging studies of working memory and intelligence, two hypotheses were made.

First, it was hypothesized that subjects with a high IQ would respond correctly more often during the spatial working memory task than subjects with an average IQ. It was further proposed that during the rehearsal period of the spatial working memory task there would be an increase in the amplitude of the SSVEP in the right prefrontal, right premotor, right posterior parietal and right occipital areas, and a decrease in the latency of the SSVEP in the right frontal area that would be greater for the high IQ group than for the average IQ group.

2. Method

2.1. Participants

The sample consisted of 25 right-handed subjects, all of whom had either normal or corrected-to-normal vision. The subjects were recruited from Swinburne University of Technology and from advertisements placed in the wider community. None of the subjects had a personal or family history of epilepsy and no personal history of neurological disorders or a serious head injury with concussion in the past 5 years.

The sample was split in half according to their full score IQ (FSIQ) on the (WAIS-R), with one subject omitted to ensure the sample size of each group was equal. The subject removed was selected to maximize the difference between the mean FSIQ scores of the two groups. The average IQ group consisted of three males and nine females (mean = 20.5 years, S.D. = 1.0 year) whose FSIQ ranged from 98 to 108 (mean = 102.8, S.D. = 4.1). The five male and seven female subjects (mean = 21.4 years, S.D. = 2.6 years) who comprised the high IQ group had a FSIQ ranging from 109 to 135 (mean = 122.9, S.D. = 9.2).

The study was approved by the Swinburne University Human Research and Ethics Committee and written informed consent was provided by all subjects.

2.2. Materials

2.2.1. Psychometric tests

The WAIS-R (Wechsler, 1981) was used to assess psychometric intelligence. The WAIS-R is an individually administered test that consists of 11 subtests that provide a verbal IQ, a performance IQ, and a full scale IQ. The WAIS-R has been found to be internally consistent, with the split-half reliability of the FSIQ ranging between .96 and .98 (Wechsler, 1981).

The handedness of each subject was assessed using the Edinburgh Handedness Inventory (Oldfield, 1971), which is a questionnaire requiring the subject to mark their preferred hand when completing 12 standard tasks. The tasks include writing, opening the lid of a box and using scissors.

Subjects were given a pre-recording questionnaire inquiring about demographic variables such as age, gender and education, and the presence of any personal or family history of epilepsy and neurological disorders.

Subjects were also given a post-recording questionnaire to assess the strategies used to perform the spatial working memory task. The questionnaire contained questions concerning the difficulty of the task, where their gaze was focused during the task, whether performance improved with practice, and the strategies used to perform the tasks.

2.2.2. Working memory tasks

The computerized spatial working memory task was a variation of the paradigm used by Jonides et al. (1993). The subjects were required to fixate on a cross, which remained on the screen for the duration of the task, a set of dots would then appear on an imaginary circle around the cross for 300 ms before disappearing. The subjects were required to remember the location of the dots for approximately 3 s until a probe circle appeared, then to decide if the circle covered any of the previously presented dots. Subjects responded by pressing either the yes (right hand) or the no buttons (left hand).

There was both a high and low demand spatial working memory task, with the high demand task involving three dots appearing in any of 24 possible locations on an imaginary circle. The low demand spatial working memory task required the subjects to retain the locations of the two dots, which could only appear in one of four positions on the imaginary circle: 12, 3, 6 and 9 o'clock. The task protocol is shown in Fig. 1.

[FIGURE 1 OMITTED]

The high and low demand tasks were pseudo-randomly grouped into three blocks of trials that lasted for approximately 4 min each. Each block of trials consisted of 20 high and 20 low demand spatial working memory tasks, and the timing of the tasks was controlled to millisecond accuracy by computer.

The subjects also performed the continuous performance task (CPT), a task assessing simple visual vigilance. It involves the repeated appearance of the numbers 1 to 5 in sequence for 4 min. The subjects were required to observe the numbers and press both the left and right buttons simultaneously whenever the number 5 appeared on the screen.

The CPT was included as an external control task as it has a very low working memory demand. It also required a lower level of sustained attention than the working memory tasks, as the subjects knew when the stimuli were going to appear and what the stimuli would be, unlike the working memory tasks where the subject did not know whether they were going to have to remember two or three locations.

2.2.3. Stimulus

The display presenting the dots subtended an angle of 3.1 [degrees] both vertically and horizontally when viewed by the subjects at a fixed distance of 1.3 m. Each of the numbers in the CPT subtended a horizontal angle of 0.9 [degrees], and a vertical angle of 1.1 [degrees]

The SSVEP was evoked using a 13-Hz sinusoidal flickering red light that was reflected over the subjects' visual field using two light-emitting diode arrays attached to a pair of goggles with half-silvered mirrors. The flickering light subtended a horizontal angle of 160 [degrees] and a vertical angle of 90 [degrees]. When viewed against the background the difference between the maximum and minimum luminance (modulation depth) of the flicker was 45%.

2.2.4. Recording equipment

The EEG was recorded from 64 sites on the scalp, which included all the International 1020 recording sites. The recording sites are shown in Fig. 2, and the exact locations of the recording sites have been described by Silberstein et al. (1990). A helmet that was designed and built at the Swinburne Brain Sciences Institute (Ciorciari, Silberstein, Simpson, & Schier, 1987) held the electrodes in place. An IBM 486 compatible computer was used for the control of the stimulus presentation, data acquisition and signal processing.

[FIGURE 2 OMITTED]

2.3. Procedure

2.3.1. Psychometric testing

The WAIS-R was administered to each subject individually according to the standardized testing procedure (Wechsler, 1981).

Just prior to the SSVEP recording session, the subjects completed the Edinburgh Handedness Inventory (Oldfield, 1971) and prerecording questionnaire to confirm that the subject had not consumed caffeine or nicotine 2 h prior to testing, and had refrained from drinking alcohol 24 h prior to testing.

2.3.2. SSVEP recording

Subjects were seated comfortably in a darkened room approximately 1.3 m away from the computer presenting the stimuli.

Electrodes on both earlobes were used as a reference and an electrode on the nose served as a ground. The impedance of the electrodes ranged from 5-25 k [Omega] when measured at 40 Hz. The brain electrical activity was amplified 20,000 times and bandpass filtered down 3 dB at 0.1 and 30 Hz. The data was then digitized to 12-bit accuracy at a rate of 200 Hz.

Prior to recording, subjects were given instructions on how to perform the spatial working memory tasks. Subjects were then allowed to practice the tasks for approximately 2 min. During recording subjects were exposed to three 4-min blocks of spatial working memory trials with a small break between each block of trials. After the spatial working memory task, subjects were instructed on how to perform the CPT, and received approximately 30 s of practice, followed by a 4-min block of trials. A post-recording questionnaire was administered at the completion of all the conditions, assessing strategy use during the spatial working memory tasks.

2.3.3. Signal processing

The single cycle Fourier coefficients were calculated over 20 stimulus cycles using a cosine window at the 13 Hz stimulus frequency, resulting in a temporal resolution of 0.77 s. The 20 cycle evaluation period was then moved one stimulus cycle and the Fourier coefficients were recalculated. This was done until the entire 246 s of EEG data for each task was analyzed, and then repeated for the data recorded from each of the 64 electrode sites. A more detailed description of the SSVEP signal processing has been reported by Silberstein et al. (1990).

To assess the changes in the SSVEP while the positions of the dots were being held in working memory, 10-s epochs of the SSVEP real and imaginary components centered on the disappearance of the target were averaged for both the high and low demand spatial working memory tasks. Similarly, 10-s epochs of data were averaged for the CPT, with the epoch centered on the appearance of the number 2. This procedure was repeated for each of the 64 electrodes.

2.3.4. Cross-subject averaging of the SSVEP

Prior to averaging across the subjects, the SSVEP amplitude for each individual was normalized. This was done by calculating the mean SSVEP amplitude across all Fourier coefficients corresponding to the 246 s of the CPT for each of the 64 electrodes separately. These mean amplitude values were then averaged together to produce a normalization factor for each subject. This normalization factor was then divided into all the averaged epoch data. The normalized SSVEP epoch data was then averaged for the subjects in the average and high IQ groups separately, and differences in amplitude and phase between the working memory task and the CPT were calculated for the two groups.

2.3.5. The subtraction method and topographical mapping of the SSVEP

Topographic maps of the difference in amplitude and phase between the working memory and CPT tasks were produced for the average and high IQ groups. This was achieved by subtracting the SSVEP phase and normalized amplitude of both the spatial working memory tasks from the SSVEP phase and normalized amplitude of the CPT for the two groups separately. In addition, the phase and normalized amplitude of the high demand task was subtracted from the phase and normalized amplitude of the low demand task for the two groups separately. The phase (in radians) was then converted to latency (ms), by multiplying the phase by -- 12.2427. (1)

The subtraction method has been previously employed in imaging studies (Gold et al., 1996; Jonides et al., 1993; Petrides et al., 1993; Smith et al., 1995). To investigate the topographic variations in the SSVEP amplitude and latency associated with spatial working memory, the CPT was used as a control task. Comparing the working memory tasks and the CPT should result in the subtraction of changes in the SSVEP that are common to both, such as aspects of attention and vigilance (Smith & Jonides, 1997). Therefore, the subtraction method should reveal the brain electrical activity underlying the working memory processes involved in maintaining the spatial information.

To investigate the differences in the SSVEP amplitude and latency associated with increasing the memory demands on spatial working memory processes, the low demand spatial working memory task was compared to the high demand spatial working memory task. The low and the high demand working memory tasks required similar levels of attention and vigilance, the only difference between the two tasks were having to retain two locations in the low demand task versus three locations in the high demand task. Therefore, the cortical activity associated with attention and vigilance that was common to both the tasks should be subtracted out, revealing the effects of increasing the working memory load on the SSVEP.

Topographical maps of Hotelling's T were used to display the statistical strength of the SSVEP amplitude and latency changes. The Hotelling's T values were derived from multiple bivariate paired [T.sup.2] tests, in which the SSVEP at each of the 64 electrodes for the holding period (after the dots have disappeared) of both the high and low demand spatial working memory tasks were compared to the mean SSVEP response centered on the disappearance of the number 2 in the CPT. Bivariate tests were used, as the SSVEP consisted of complex numbers that are expressed as both real and imaginary components (Silberstein et al., 1995).

In the regions of the Hotelling's T maps where T (the positive square root of [T.sup.2]) was equal to 2.20, 3.11, 3.35 and 4.43, white contour lines were added that corresponded to the significance levels of 5%, 1%, 0.5% and 0.1%, respectively, for a single comparison. The current study made multiple, rather than a single comparison, so the P values needed to be adjusted to take into account the increased risk of making a Type I error.

It has been suggested that using a Bonferroni correction of .05/64 to take into account the 64 recording sites is overly conservative as it overlooks the high correlations between neighboring recording sites (Duffy et al., 1990; Nunez, 1981; Silberstein & Cadusch, 1992). A technique known as spatial principal components analysis has been used to estimate the spatial dimensionality of spontaneous EEG and evoked potential data. Studies have indicated that no more than five factors are needed to account for more than 95% of the variance sites (Duffy et al., 1990; Silberstein & Cadusch, 1992). Therefore, for a single set of comparisons, derived from 64 scalp recording sites, the Bonferroni adjusted P value would be 1% (i.e. .05/5). For the purposes of the current study, the P value then needs to be further adjusted because six Hotelling's [T.sup.2] maps (2 IQ groups x 3 comparisons) were used to display the findings. Therefore, after taking into account the multiple comparisons, a Bonferroni-corrected P value of. 167% is needed for a maximum probability of falsely rejecting the null hypothesis of 5%.

3. Results

To examine whether the high demand spatial working memory task was more difficult than the low demand spatial working memory task, the mean percentage of correct responses for both the high and low demand tasks were calculated. Subjects were found to be correct more often in the low demand task (mean = 88.5%, S.D. = 4.0%) than in the high demand task (mean = 62.1%, S.D. = 6.7%). A paired-samples t test revealed that this difference was significant (t(24) = 21.65, P < .001), indicating that the low demand spatial working memory task was easier than the high demand task.

It was predicted that the high IQ group would perform better on the working memory task than the average IQ group; however, comparing the mean percentage correct for the average IQ group and the high IQ group on the low demand spatial working memory task (mean = 88.8%, S.D. = 4.1%; mean = 88.8%, S.D. = 3.8%), and on the high demand spatial working memory task (mean = 60.9%, S.D. = 7.0%; mean = 63.4%, S.D. = 6.7%) revealed no significant difference between the performance of the two groups. However, one-tailed bivariate correlations revealed a moderate strength, positive correlation between FSIQ and percentage correct on the high demand working memory task when corrected for the restricted range of FSIQ scores (r = .40, N = 24, P < .05; uncorrected r = .34, N = 24, NS), but no relationship with the low demand working memory task (r = .22, N = 24, NS; uncorrected r = .20, N = 24, NS). This indicated that there was a tendency for individuals with a higher FSIQ to perform better on the high demand working memory task than those with a lower FSIQ. Although a t test revealed that the correlations between FSIQ and performance on the high and low demand working memory tasks were not significantly different (t = .88, df = 21, NS).

To investigate if the SSVEP amplitude and latency during the working memory tasks was influenced by the subjects' level of intelligence, the SSVEP recorded during the `holding' period of the working memory tasks was compared to the SSVEP recorded during the visual vigilance task for the average and high IQ groups separately, and are displayed in Figs. 3 and 4. The warmer colors represent a reduction in the amplitude and a decrease in the latency of the S SVEP during the working memory tasks compared to the CPT.

Generally, it was found that SSVEP amplitude and latency differences were similar when contrasting both the low (Fig. 3) and high (Fig. 4) demand working memory tasks with the CPT. SSVEP differences appeared to be larger when comparing amplitude and latency topographies between the two IQ groups.

[FIGURE 3 AND 4 OMITTED]

During the low demand spatial working memory task the high IQ group revealed an increase in SSVEP latency in the frontal area compared to the SSVEP during the CPT. This difference was also found in the average IQ group, but it appeared to be smaller and more diffuse than for the high IQ group. The results also revealed that the high IQ group, but not the average IQ group, showed an increase in SSVEP amplitude in the left occipital and parietal areas and a decrease in latency bilaterally in the occipital and parietal areas during the low demand working memory task compared to the CPT. Although, as was shown in the low demand working memory task Hotelling's T maps (Fig. 3), only the SSVEP differences in the frontal area for the high IQ group reached the 5% significance level after correcting for the number of comparisons (T > 4.43, P < .001), where the SSVEP differences in the frontal area for the average IQ group just failed to reach significance (T > 3.35, NS).

A similar pattern was evident during the high demand spatial working memory task (Fig. 4). Although for the high IQ group there was a greater decrease in SSVEP latency in the occipital and parietal areas. The Hotelling's T maps shown in Fig. 4 indicated that the SSVEP differences between the high demand working memory task and the CPT in the occipital area for the high IQ group just failed to reach the 5% significance level after correcting for the number of comparisons (T > 3.35, NS). Contrary to the Hotelling's T maps for the low demand working memory task, the Hotelling's T maps for the high demand working memory task did not reveal any significant SSVEP differences between the high demand working memory task and the CPT, with the SSVEP differences in the frontal area for both the average and high IQ groups just failing to reach significance after correcting for the number of comparisons (T > 3.35, NS), indicating that the SSVEP amplitude and latency during the high demand working memory task was not significantly different from that during the CPT.

To assess the effect of increasing memory load on the brain electrical activity, the SSVEP amplitude and latency during the high and low demand working memory tasks were compared for each IQ group, and are displayed in Fig. 5.

[FIGURE 5 OMITTED]

The differences between the SSVEP amplitude and latency during the high and low demand spatial working memory task for both the average and high IQ groups were much smaller than those found by subtracting the SSVEP recorded during the spatial working memory tasks from the SSVEP recorded during the CPT. The topography of the SSVEP was also different, with the high demand working memory task eliciting a larger SSVEP latency attenuation in the right temporal areas than during the low demand working memory task, for both IQ groups. There were also some apparent differences in the SSVEP topography between the two IQ groups, with the high IQ group showing a reduction in SSVEP amplitude in the right occipital area, and a decrease in the SSVEP latency in the occipital and parietal areas during the high demand working memory task, compared to the low demand working memory task. Although, as can be seen in the Hotelling's T maps (Fig. 5), the differences in the SSVEP between the high and low demand working memory tasks for both IQ groups were not significantly different.

4. Discussion

The current study did not find a significant difference in working memory performance between the average and high IQ groups, although it was found that subjects with higher FSIQ scores performed better than subjects with lower FSIQ scores on the high demand spatial working memory. The current study also found evidence that the brain electrical activity of the two groups differed during the working memory tasks.

The moderate strength, positive relationship between performance on the high demand working memory task and FSIQ was consistent with previous studies reporting a positive relationship between reasoning ability and intelligence and working memory capacity (Embretson, 1995; Kyllonen & Christal, 1990; Necka, 1992). The lack of a significant relationship between performance on the low demand working memory task and FSIQ in the current study may have been due to a ceiling effect, as the subjects in the current study reported that the low demand working memory task was very simple, and had an accuracy rate of nearly 90%. It may have also been due to the type of working memory task used. For example, several studies have found that simple encoding paradigms, such as the one used in the current study, may be a poorer predictor of intelligence than more complex tasks that not only require encoding, but also the rapid manipulation of information in working memory (Larson, Merritt, & Williams, 1988; Larson & Saccuzzo, 1989).

This notion was further supported by evidence that the correlational studies investigating the relationship between working memory and intelligence that reported only a weak-to-moderate strength relationship involved simple `delayed matching-to-sample' tasks (Necka, 1992). Previous studies that have reported a strong relationship between working memory and intelligence have used more complex tasks such as mental arithmetic and backward digit span (Embretson, 1995; Kyllonen & Christal, 1990).

The relationship between intelligence and working memory was also evident in the electrical activity of the brain during the working memory tasks. The topography and the magnitude of the SSVEP amplitude and latency differences during the working memory tasks differed between the average and high IQ groups.

The major finding of the current study was that during the period that spatial working memory was engaged in the rehearsal of information, there was an increase in the SSVEP latency frontally, and a decrease in the SSVEP latency and an increase in SSVEP amplitude posteriorly that was greater for the high IQ group than the average IQ group. These results were not consistent with the hypothesis that there would be an increase in the amplitude of the SSVEP in the right prefrontal, right premotor, right posterior parietal and right occipital areas, and a decrease in the latency of the SSVEP in the right frontal area during the working memory tasks. However, they were consistent with the hypothesis that the magnitude of these changes would be greater for the high IQ group than for the average IQ group.

The lack of lateralization to the right hemisphere during the spatial working memory task was contrary to previous research that has reported that cortical activity during spatial working memory tasks is predominantly based in the right hemisphere (Jonides et al., 1996; Smith et al., 1995). Jonides et al. (1993) and Smith et al. (1995) reported significant activation in the right hemisphere during spatial working memory tasks, but also reported that there was some activation of the left hemisphere during working memory. Smith et al. attributed this finding to `spillover' from the right hemisphere because of the difficult nature of the task. Ruchkin et al. (1992), using ERPs with a shorter temporal resolution than PET, reported that the high amplitude negative slow wave that was evident over the right parietal area became less lateralized to the right hemisphere when the subject was holding the information in their working memory. Therefore, this may indicate that the proposed predominantly right hemispheric cortical activity during spatial working memory tasks may become less evident during the periods when the information is being rehearsed. This is consistent with the current study's findings of a predominately bilateral distribution of SSVEP changes during the holding period of the working memory task.

The increase in SSVEP amplitude during the high demand spatial working memory task in the parietal and occipital areas that was evident in the high IQ group was consistent with topographical findings of previous electrophysiological and neuroimaging studies (Gold et al., 1996; Jonides et al., 1993; Petrides et al., 1993; Ruchkin et al., 1992, 1995, 1997; Smith et al., 1995). The increase in the SSVEP amplitude in the posterior parietal area during the working memory tasks compared to the CPT was consistent with the recent study by Silberstein (1997), that investigated spatial working memory using the SSPT. Although, contrary to Silberstein's findings of an increase in SSVEP amplitude and a decrease in SSVEP latency in the prefrontal area, the current study found an increased SSVEP latency in the frontal area and a decreased SSVEP latency posteriorly during the working memory tasks when compared to the CPT.

The increased SSVEP amplitude while the spatial information was being retained in working memory was also consistent with the findings of Krause et al. (1996), Okada and Salenius (1998), Pfurtscheller (1992) and Ray and Cole (1985), who also reported an enhancement or increase in the amplitude of high alpha EEG activity.

Previous studies have suggested that occipital processes play a role in generating an internal image of the visual stimuli during a spatial working memory task (Jonides et al., 1993; Smith et al., 1995). Smith and Jonides (1997) and Smith et al. (1995) have proposed that the posterior parietal area may be involved in the storage and rehearsal of spatial information held in working memory. This may indicate that high intelligence is characterized at least partially by the quality of posterior processes to store and rehearse spatial information in working memory.

While the current study found no change in SSVEP amplitude in the frontal areas during the spatial working memory tasks, the results did show an increase in the SSVEP latency in frontal areas that was greater for the high IQ group than for the average IQ group. Previous spatial working memory studies have reported that the frontal areas are one of the most significant areas of activation during working memory (Gold et al., 1996; Jonides et al., 1993; McCarthy et al., 1994; Petrides et al., 1993; Ruchkin et al., 1992, 1995, 1997; Smith et al., 1995). There is a general consensus that the prefrontal area is the executive in working memory, storing information and coordinating the other components of the working memory model (Raichle, 1993; Wickelgren, 1997). Smith and Jonides (1997) and Smith et al. (1995) have also suggested that the premotor area may contribute to the spatial working memory rehearsal loop. The current study found that when comparing the high and low demand working memory tasks to the CPT, the high IQ group showed a larger latency increase in the frontal area than the average IQ group when the positions of the dots were being held in working memory. This may indicate that the quality of the frontal executive processes, as well as the quality of the posterior storage and rehearsal processes, play an important role in influencing intelligence.

When increasing the memory load of the working memory task for the high IQ group there was a small increase in SSVEP latency in the occipital, parietal and right temporal areas, and a decrease in amplitude in the right occipital area. This was consistent with the results reported in Ruchkin et al.'s (1992, 1995, 1997) ERP studies, who reported that the slow wave activity evident during the spatial working memory task was responsive to changes in working memory load, particularly in the posterior areas. However, in the present study, the average IQ group failed to show posterior changes in SSVEP amplitude and latency. This provides further support that the quality of the posterior rehearsal and storage processes is an important factor influencing intelligence.

In summary, there was a tendency for individuals with a higher IQ to perform better on the high demand working memory task than those with a lower IQ. This relationship between working memory and intelligence was also evident in SSVEP changes. The topography and magnitude of the SSVEP amplitude and latency changes differed between the two groups. While further research is required, the results of the present study indicate that when rehearsing spatial information in working memory there was an increase in SSVEP latency in frontal areas, and an increase in SSVEP amplitude and a decrease in SSVEP latency in the parietal and occipital areas. These differences were greater in magnitude for the high IQ group than for the average IQ group. Increasing the memory load of the task revealed small SSVEP amplitude and latency changes with a decreasing SSVEP amplitude in the right occipital area and a decrease in SSVEP latency posteriorly and in the right temporal area for the high IQ group. For the average IQ group, increasing the memory load revealed a small decrease in SSVEP latency in the right temporal area. These results indicate that high intelligence may be characterized by the quality of the frontal executive processes and the ability to store and rehearse spatial information in working memory in the posterior areas of the brain. Although this interpretation is a somewhat speculative result, it raises some important questions regarding the neural processes underlying working memory and how intelligence and working memory are related.

(1) This value is used to convert phase to latency as 1 radian = 1/13 / 2 [pi] seconds, therefore 1 radian = 12.2427.

References

Baddeley, A. (1988). Cognitive psychology and human memory. Trends in Neurosciences, 11, 176-181.

Baddeley, A. (1992a). Working memory. Science, 255, 556-559.

Baddeley, A. (1992b). Is working memory working? The fifteenth Bartlett lecture. Quarterly Journal of Experimental Psychology, 44A, 1-31.

Baddeley, A. (1995). Working memory. In: M. S. Gazzaniga (Ed.), The cognitive neurosciences (pp. 755-764). Cambridge, MA: MIT Press.

Baddeley, A., Logie, R., Nimmo-Smith, I., & Brereton, N. (1985). Components of fluent reading. Journal of Memory and Language, 24, 119-131.

Baddeley, A. D., & Hitch, G. (1974). Working memory. In: G. H. Bower (Ed.), The Psychology of learning and motivation vol. 8, (pp. 47-89). New York: Academic Press.

Carpenter, P. A., Just, M. A., & Shell, P. (1990). What one intelligence test measures: a theoretical account of the processing in the Raven Progressive Matrices Test. Psychological Review, 97, 404-431.

Ciorciari, J., Silberstein, R. B., Simpson, D. G., & Schier, M. A. (1987). The multi-channel electrode helmet. Paper presented at the Proceedings Conference on Engineering and Physical Sciences in Medicine, Melbourne, Australia.

Duffy, F. H., Jones, K., Bartels, P., Albert, M., McAnulty, G. B., & Als, H. (1990). Quantified neurophysiology with mapping: statistical inference, exploratory and confirmatory data analysis. Brain Topography, 3, 3-12.

Embretson, S. E. (1995). The role of working memory capacity and general control processes in intelligence. Intelligence, 20, 169-189.

Engle, R. W., Cantor, J., & Carullo, J. J. (1992). Individual differences in working memory and comprehension: a test of four hypotheses. Journal of Experimental Psychology: Learning, Memory and Cognition, 18, 972-992.

Gevins, A., Smith, M. E., McEvoy, L., & Yu, D. (1997). High-resolution EEG mapping of cortical activation related to working memory: effects of task difficulty, type of processing, and practice. Cerebral Cortex, 7, 374-385.

Gold, J. M., Berman, K. F., Randolph, C., Goldberg, T. E., & Weinberger, O. R. (1996). PET validation of a novel prefrontal task: delayed response alternation. Neuropsychology, 10, 3-10.

Haier, R. J., Siegel, B., Tang, G., Abel, L., & Buchsbaum, M. S. (1992). Intelligence and changes in regional cerebral glucose metabolic rate following learning. Intelligence, 16, 415-426.

Haier, R. J., Siegel, B. V., Nuechterlein, K. H., Hazlett, E., Wu, J. C., Paek, J., Browning, H. L., & Buchsbaum, M. S. (1988). Cortical glucose metabolic rate correlates of abstract reasoning and attention studied with positron emission tomography. Intelligence, 12, 199- 217.

Jonides, J., Reuter-Lorenz, P. A., Smith, E. E., Awh, E., Barnes, L. L., Drain, M., Glass, J., Lauber, E. J., Patalano, A. L., & Schumacher, E. H. (1996). Verbal and spatial working memory in humans. In: D. L. Medin (Ed.), The psychology of learning and motivation vol. 35, (pp. 43-88). San Diego, CA: Academic Press.

Jonides, J., Smith, E. E., Koeppe, R. A., Awh, E., Minoshima, S., & Mintun, M. A. (1993). Spatial working memory in humans as revealed by PET. Nature, 363, 623-625.

Just, M. A., & Carpenter, P. A. (1992). A capacity theory of comprehension: individual differences in working memory. Psychological Review, 99, 122-149.

Krause, C. M., Lang, A. H., Laine, M., Kuusisto, M., & Porn, B. (1996). Event-related EEG desynchronization and synchronization during an auditory memory task. Electroencephalography and Clinical Neurophysiology, 98, 319-326.

Kyllonen, P. C., & Christal, R. E. (1990). Reasoning ability is (little more than) working-memory capacity?! Intelligence, 14, 389-433.

Larson, G. E., Merritt, C. R., & Williams, S. E. (1988). Information processing and intelligence: some implications of task complexity. Intelligence, 12, 131-147.

Larson, G. E., & Saccuzzo, D. P. (1989). Cognitive correlates of general intelligence: towards a process theory of g. Intelligence, 13, 5-31.

McCarthy, G., Blamire, A. M., Puce, A., Nobre, A. C., Bloch, G., Hyder, F., Goldman-Rakic, P., & Shulman, R. G. (1994). Functional magnetic resonance imaging of human prefrontal cortex activation during a spatial working memory task. Proceedings of the National Academy of Sciences of the United States of America, 91, 8690-8694.

Necka, E. (1992). Cognitive analysis of intelligence: the significance of working memory processes. Personality and Individual Differences, 13, 1031 - 1046.

Nunez, P. (1981). Electric fields of the brain: the neurophysics of the EEG. New York: Oxford Univ. Press.

Okada, Y. C., & Salenuis, S. (1998). Roles of attention, memory and motor preparation in modulating human brain activity in a spatial working memory task. Cerebral Cortex, 8, 80-96.

Oldfield, R. C. (1971). The assessment and analysis of handedness: the Edinburgh Inventory. Neuropsychologia, 9, 73-113.

Papanicolaou, A. C., & Johnstone, J. (1984). Probe evoked potentials: theory, method and applications. International Journal of Neuroscience, 24, 107-131.

Pellegrino, J. W., & Glaser, R. (1979). Cognitive correlates and components in the analysis of individual differences. Intelligence, 3, 187-214.

Petrides, M., Alivisatos, B., Evans, A. C., & Meyer, E. (1993). Dissociation of human mid-dorsolateral from posterior dorsolateral frontal cortex in memory processing. Proceedings of the National Academy of Sciences of the United States of America, 90, 873-877.

Pfurtschellar, G. (1992). Event-related synchronisation (ERS): an electrophysiological correlate of cortical areas at rest. Electroencephalography and Clinical Neurophysiology, 83, 62-69.

Prabhakaran, V., Smith, J. A., Desmond, J. E., Glover, G. H., & Gabrieli, J. D. E. (1997). Neural substrates of fluid reasoning: an fMRI study of neocortical activation during performance of the Raven's progressive matrices. Cognitive Psychology, 33, 43-63.

Raichle, M. E. (1993). The scratchpad of the mind. Nature, 363, 583-584.

Ray, W. J., & Cole, H. W. (1985). EEG alpha activity reflects attentional demands and beta activity reflects emotional and cognitive processes. Science, 228, 750-752.

Regan, D. (1989). Human brain electrophysiology: evoked potentials and evoked magnetic fields in science and medicine. New York: Elsevier.

Ruchkin, D. S., Canoune, H., Johnson, R., & Ritter, W. (1995). Working memory and preparation elicit different patterns of slow wave event-related brain potentials. Psychophysiology, 32, 399-410.

Ruchkin, D. S., Johnson, R., Grafman, J., Canoune, H., & Ritter, W. (1992). Distinctions and similarities among working memory processes: an event-related potential study. Cognitive Brain Research, 1, 53-66.

Ruchkin, D. S., Johnson, R., Grafman, J., Canoune, H., & Ritter, W. (1997). Multiple visuospatial working memory buffers: evidence from spatiotemporal patterns of brain activity. Neuropsychologia, 35, 195-209.

Silberstein, R. B. (1995). Neuromodulation of neocortical dynamics. In: P. Nunez (Ed.), Neocortical dynamics and human EEG rhythm (pp. 591-627). New York: Oxford Univ. Press.

Silberstein, R. B. (1997). The steady-state visually evoked, neocortical dynamics and cognitive function. Paper presented at the Proceedings of the 3rd Pan Pacific Conference on Brain Topography, Tokyo, Japan. (April).

Silberstein, R. B., & Cadusch, P. J. (1992). Measurement processes and spatial principal components analysis. Brain Topography, 4, 267-276.

Silberstein, R. B., Ciorciari, J., & Pipingas, A. (1995). Steady-state visually evoked potential topography during the Wisconsin card sorting test. Electroencephalography and Clinical Neurophysiology, 96, 24-35.

Silberstein, R. B., Schier, M. A., Pipingas, A., Ciorciari, J., Wood, S. R., & Simpson, D. G. (1990). Steady-state visually evoked potential topography associated with a visual vigilance task. Brain Topography, 3, 337-347.

Smith, E. E., & Jonides, J. (1997). Working memory: a view form neuroimaging. Cognitive Psychology, 33, 5-42.

Smith, E. E., Jonides, J., Koeppe, R. A., Awh, E., Schumacher, E. H., & Minoshima, S. (1995). Spatial versus object working memory: PET investigations. Journal of Cognitive Neuroscience, 7, 337-356.

Smith, M. E., Rush, G., & Gevins, A. (1998, July). Neuroelectric signals of working memory predict IQ test scores. Poster presented at the XIIth International Conference on Event-Related Potentials of the Brain, Cambridge, MA.

Swanson, H. L., & Berninger, V. (1995). The role of working memory in skilled and less skilled readers' comprehension. Intelligence, 21, 83-108.

Wechsler, D. (1981). Wechsler adult intelligence scale-revised manual. New York: Psychological Corporation.

Wickelgren, I. (1997). Working memory linked to intelligence. Science, 275, 1581.

Wilson, G. F., & O'Donnell, R. O. (1986). Steady-state evoked responses: correlations with human cognition. Psychophysiology, 23, 57-61.

C. Van Rooy (a), C. Stough (a, b) *, A. Pipingas (a), C. Hocking (a), R.B. Silberstein (a)

* Corresponding author. Tel.: +61-3-9214-8167; fax: +61-3-9214-5230. E-mail address: cstough@swin.edu.au (C. Stough).

(a) Swinburne University of Technology, Brain Sciences Institute, PO. Box 218, Hawthorn, Victoria 3122, Australia (b) Swinburne University of Technology, Neuropsychology Laboratory, PO. Box 218, Hawthorn, Victoria 3122, Australia

Received 30 November 1998; received in revised form 6 October 1999; accepted 6 October 1999
COPYRIGHT 2001 Ablex Publishing Corp.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2001 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Van Rooy, C.; Stough, C.; Pipingas, A.; Hocking, C.; Silberstein, R.B.
Publication:Intelligence
Geographic Code:1USA
Date:Oct 1, 2001
Words:8380
Previous Article:Item sequencing effects on the measurement of fluid intelligence.
Next Article:The relationship between the ability to divide attention and standard measures of general cognitive abilities.
Topics:

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters