EEG Gamma Band Is Asymmetrically Activated by Location and Shape Memory Tasks in Humans

Abstract

From the viewpoint of psychology, it is thought that perception analysis of the visual world includes two information processes: global (whole) and local (part) processes. It is assumed that the global process is carried out in the right hemisphere, and the local process, in the left hemisphere. In the present study, gamma EEG band activities during location memory (LM) task, as a global form, and shape memory (SM) task, as a local form, were calculated from the temporal, parietal and occipital areas using stimuli consisting of categorical patterns of small shapes. Gamma band activity during the SM task was greater than that during the LM task. It was assumed that the SM task requires a higher memory load condition than the LM task. In terms of the laterality ratio obtained from the whole electrode array, the gamma band was significantly activated in the right hemisphere during the LM task, and in the left hemisphere during the SM task. The gamma activation in the occipital area was significantly high in the right hemisphere for both tasks. High gamma band activation was observed in the right parietal area during the LM task and in the left temporal area during the SM task. It was concluded that global and local information processes occur in the left temporal areas and in the right occipitoparietal areas, respectively. The results of this study are useful in the assessment of visual cognition deficits in patients with cerebral hemispheric lesions in the physical therapy.

It is well known in primates that the visual information process is carried out via two cortical pathways: the occipitoparietal pathway, related to visuospatial recognition, and the occipitotemporal pathway, related to object recognition^1–3). The presence of the two pathways has also been suggested in humans in neuropsychological studies of patients⁴⁾⁵⁾ and by positron emission tomography (PET) in normal subjects^6–8). Furthermore, it is known that brain function is related to attention, cognition and memory of visual spaces and objects are asymmetrically located in the left and right hemispheres^9–11).

In the field of neuropsychology, Navon et al.¹²⁾ and Shulman et al.¹³⁾ pointed out that cognition of visual subjects is accomplished by two information process systems: global and local processes. In other words, a visual scene consists of global and local forms, and humans analyze its characteristics by sorting them into to global and local information process systems. Based on behavioral studies, it is assumed that the global process is localized in the right hemisphere, and the local process, in the left hemisphere^14–17). Fink et al.¹⁸⁾ employed PET to examine brain activity during character cognitive task using hierarchical stimuli consisting of large characters (as a global form), and small characters (as a local form). They showed that the cognitive task of the global form activated the right prestriate cortex, while the cognitive task of the local form activated the left inferior occipital cortex. Their results seem to support the general concept of functional asymmetry of the human brain. However, Fink et al.¹⁰⁾ recently reported contradictory results using hierarchical stimuli consisting of figures rather than characters. Thus, the functional localization of the two information process systems is not yet neurophysiologically understood.

Gamma band EEG activity (usually in the 30 to 60 Hz band range) is proposed to be the frequency band which reflects cortical activity related to cognitive processes¹⁹⁾²⁰⁾ as well as short-term memory in visual discrimination task²¹⁾. Numerous reports have shown that gamma band activity is a useful tool to study functional activities of the hemispheres in visual information processes.

More detailed information is needed for evaluation for visual recognition in patients with cerebral vascular accident on physical therapy. In the present study, therefore, gamma band activities during location memory (LM) task, as a global form, and shape memory (SM) task, as a local form, were calculated from EEGs recorded from temporal, parietal and occipital areas. For the LM and SM tasks, stimuli consisting of categorized patterns of small shapes were used to reveal functional localization of global and local memory processes in the hemispheres.

Methods

Subjects

Eight right-handed college students (6 males and 2 females, 19 to 27 years old) participated in the present study. All were free of neurological disorders.

Visual memory tasks

Figure 1 shows the flow chart of stimulation used for LM and SM tasks. One stimulation epoch was initiated by displaying a fixation point (cross) at the center of CRT at random intervals between 3,000 and 5,000 ms. Five types of figures (circle, square, upward-and downward-triangles and cross) were displayed for 60 ms. Immediately after the onset of the fixation point display, three out of the five types of figures (as a memory display), and at 2,000 ms after the onset, one out of the five types of figures (as a recall display) were presented at the center of an imaginary matrix of 3 × 3 squares on the CRT display. The center of the innermost square contained the fixation point, while the figures for memory or recall displays were presented at the center of one of the 8 outer squares (visual angle of each square: 0.9° and 1.3°). The locations for memory display were randomly arranged by a computer software.

Fig. 1.

Schematic illustration of the events in the location memory (LM) and shape memory (SM) tasks. In both tasks, subjects were asked to fixate a cross in centre of the CRT, and then memory and recall displays appeared for 60 ms respectively.

During the experiments, the subjects were instructed to gaze at the fixation point. In the LM task, the subjects were asked to memorize the position of the displayed figure (S1) (disregarding its shape) and to discriminate whether the location of the recall figure (S2) matches that of S1. In the SM task, they were asked to memorize the shape of the display figure (S1) (disregarding its location) and to discriminate whether the shape of the recall figure (S2) matches that of S1. The above two tasks were arranged so that each task consisted of 2 blocks, and each block consisted of 30 epochs. Four blocks in the 2 tasks were pseudorandomly carried out so that the frequency of correct answers became 30% for each task. After showing a recall display, the subjects were instructed to quickly press a button with their right index finger if S2 matched S1 in either task.

EEG recording and quantification

The EEG data was recorded from 21 electrodes affixed to the scalp according to the international 10–20 system with reference to linked ears. EEG data of 5,000 ms duration [500 ms before and 4,500 ms after onset of memory display (S1)] were amplified with a band pass filter of 0.05–100 Hz and digitized at a sampling rate of 200 Hz. The EOG was simultaneously recorded to facilitate rejection of EEG data contaminated by eye movements. Electrode resistance was kept at less than 5Ω.

The Fast Fourier Transform (FFT) with a Humming window was used to calculate power density (µV²/Hz) of the gamma frequency band at P₃ (left parietal), P₄ (right parietal), T₃ (left temporal), T₄ (right temporal), O₁ (left occipital), and O₂ (right occipital) areas. The evolutionary spectrum in each epoch was calculated using a time window of 320 ms, which was shifted by steps of 125 ms from the onset of the memory display to 2,000 ms afterward. The evolutionary spectrum was also calculated from EEG of 500 ms duration before the onset of the memory display (S1) as a baseline power. Gamma band activity was defined as the mean power of 40.6, 43.7 and 46.9 Hz. The mean powers at 17 latency steps during the 2,000 ms period from the onset of the memory display were subtracted from the baseline power.

Statistics

The baseline-corrected values were analyzed using the three-way ANOVA (task × electrode × 17 latency steps). The laterality ratios of the gamma band activity were also calculated at the temporal (T₃–T₄), parietal (P₃–P₄) and occipital areas (O₁–O₂) using (R−L)/(R+L) [27] and analyzed by the two-way ANOVA (task × area). A positive laterality ratio indicates a shift from the left to the right hemisphere and a negative ratio, a shift from the right to the left hemisphere. In addition, inter-group comparison was performed by the paired t-test.

Results

Three-way ANOVA (task × electrode × latency step) revealed two main effects. First, the effect of task [F(1,4692) = 4.10, p = 0.0428] was revealed due to the fact that the gamma activity of the SM task (2.08 × 10⁻⁴ µV²/Hz) was higher than that of the LM task (1.76 × 10⁻⁴ µV²/Hz). Secondly, the main effect was found due to the electrode [F (5,4692) = 18.77, p<0.00l]. The latency step did not have a main effect [F (16,112) = 0.397, p = 0.9837]. There was a significant interaction between task and electrode [F (5, 4692)= 9.149, p<0.00l]. The gamma band activity during the SM task was significantly higher than the LM task in the left temporal area (T₃), whereas that during the LM task was significantly higher than the SM task in the right parietal area (P₄) (Table 1). Two-way ANOVA of the laterality ratio (task × electrode) resulted in significant main effects of task [F (1,96) = 47.488, p<0.001]. The laterality ratio during the LM task was shifted to the right hemisphere (0.132), while that during the SM task was shifted to the left hemisphere (−0.069). Another main effect was in the electrode [F (2,96) = 106.736, p<0.00l]. The laterality ratio at the temporal area (−0.268) was shifted to that of the left hemisphere, while those of the occipital (0.157) and parietal (0.205) areas were shifted to those in the right hemisphere. There was a significant interaction between the task and electrode position [F(2,96) = 33.434, p<0.00l]. The paired t-test revealed that the laterality ratio in the LM task was shifted to the right parietal area, whereas that in the SM task was shifted to the left temporal and occipital areas (Fig. 2). The paired t-test also revealed that the reaction time during the LM task (482.9 ms) was significantly shorter than that during the SM task (552.6 ms) (d.f. (7), t=−2.415, p<0.05).

Table 1. Baseline corrected total mean gamma powers for location memory (LM) task and shape memory (SM) task (V²/Hz) × 10⁻⁴.

Fig. 2.

Mean laterality ratios (R-L)/(R+L) for location memory (LM) and shape memory (SM) tasks at temporal, parietal and occipital areas. The positive laterality ratio means a shift to the right hemisphere and negative ratio, a shift to the left hemisphere. a; p<0.001, b; p<0.0001.

Discussion

The latency range had no effect on gamma band activity in either task for 2,000 ms after the onset of memory display. Therefore, it was assumed that, during this period, cortical areas are continuously activated by cognitive processes and memory accesses imposed by the present tasks. Pasclis et al.²²⁾ reported that tasks with a high memory load condition elicit gamma band activity with a higher amplitude than those with a low memory load condition. Since it was found in the present study that the gamma band activity during the SM task is higher than that during LM task, it was assumed that the high gamma activity in the SM task is attributable to the participation of the processes related to the memory of figure shapes as well as to the memory of the location. Two memory processes were thought to be recruited in the SM task, therefore a higher memory load was imposed in the SM task than in the LM task. The above-mentioned assumption was also inferred from the fact that the SM task required a longer reaction time than the LM task.

The reaction time study using object-based hierarchical stimuli showed that presentation of an object in the left visual hemifield, that is projection to the right visual cortex, has an advantage in the discrimination of global forms, and that in the right visual hemifield (projection to the left visual cortex) has an advantage in the discrimination of local forms^15–17). These facts seem to support the idea that the global and local information processes are functionally localized in the right and left hemispheres, respectively. The present results which indicated that the laterality ratio during the LM task was observed in the right hemisphere, and that during the SM task was observed in the left hemisphere, would support the concept of such a differentiation in the functional localization between the two information processes.

The present study examined the functional localization of two tasks in the hemispheres, especially in the prestriate cortex. PET study revealed that the parietal region was bilaterally activated in spatial tasks, and that the temporal region was bilaterally activated in shape tasks^7)8)23). However, Jonides et al.¹¹⁾ reported that the activities in the right parietal and occipital areas (including the right frontal area) are increased during spatial working memory task, which agrees with the present study in that the gamma band was more significantly activated in the right parietal and occipital areas. These facts seem to suggest that the strategy for cognition and memory during LM task is to comprehend the stimuli not as a figure pattern but as a spatial location. Thus, it was assumed from the observation in the LM task that the right occipitoparietal lobe contributes to the global process. On the other hand, it was observed in the present study that the gamma band during the SM task was significantly activated in the left temporal area.

Experiments with primates have shown that the temporal lobe has important functions in visual discrimination and learning complicated figures¹⁾²⁴⁾. Furthermore, William et al.⁵⁾ reported that the function of the right temporal lobe is associated with cognition of figure patterns and faces. These facts are often inferred in relation to the observations that patients with a right hemisphere lesion draw distorted figures in the reproduction task²⁵⁾, and that patients with a left hemisphere lesion often omit details. The latter observation is assumed to indicate the relationship of the left hemisphere with local information processing. Thus, the results obtained from the SM task in the present study seemed to suggest that the activity in the left temporal area contributes to the local information process.

The gamma band in the right occipital area was more significantly activated during the SM task than during the LM task. The occipital area is considered to be important in creating a visual image²⁶⁾. Heilman et al.²⁷⁾ proposed that the left hemisphere pays attention to the right visual hemifield, while the right hemisphere pays attention to both the right and left visual hemifields. Corbetta et al.⁶⁾ reported that object-based tasks increase the activity in the right inferior occipital area, and concluded that this area plays a role in the selective attention to an object as a short-term memory system. Furthermore, Fink et al.¹⁰⁾ also assumed that the right occipital lobe has an additional function in the local information process. These proposals are in agreement with the present findings indicating that the gamma band in the right occipital area is more significantly activated during the SM task than during the LM task.

Using letter-based hierarchical stimuli and object-based hierarchical stimuli, Fink et al.⁸⁾¹⁰⁾ proposed functional localization for the global and local information processes which is in direct contrast to the prevailing theory, including that of the present study. It is thought that this discrepancy is due to the differing contents of stimuli and tasks among researchers, and that humans naturally select the best rationalized information process from among the various types of available visual processes.

The results of this study are useful in the assessment of visual cognition deficits in patients with cerebral hemispheric lesions in the physical therapy.