Education Level is Associated with Specific N200 and P300 Profiles Reflecting Higher Cognitive Functioning

Background: While different colors are commonly used during neuropsychological assessments, there is a general lack of information about the influence of education on color processing, which could affect the results of neuropsychological testing. Higher education is directly related to higher cognitive functions. Therefore, we investigated educational influences on color processing in different tasks using reaction times (RTs) in an event-related potential (ERP) study. Methods: A 128-sensor net was used for the ERP study with participants from low (G1), medium (G2) and high (G3) educational groups (n=17 per group). Participants were asked by pressing button to indicate whether they ‘liked’ or ‘disliked’ colors displayed to them for consideration in the Original Research Article Hasan et al.; JAMPS, 10(4): xxx-xxx, 2016; Article no.JAMPS.29783 2 RT analysis. A 10-20 system (19 electrode channels) was used to analyze the amplitudes and latencies of the N200 and P300 ERP components. Results: The mean differences for the like and dislike choices were calculated in terms of the amplitudes and latencies of the both components. RTs were significantly shortest in G3, then G2, and G1 reflected significantly longer RTs. Seven (out of 19) electrode locations clearly expressed N200 and P300 components. The G3 evoked the highest amplitudes (significant at T6) of both components at most of the electrode locations. The next highest amplitudes were in G2 and then G1. There was a trend toward the shortest latencies of both components being represented in G3, with G1 holding the longest latencies at most electrode sites, but this did not quite reach significance. Conclusion: Faster RTs, higher amplitudes and shorter latencies of the N200 and P300 ERP components in G3 suggest that higher education improves attention, enables faster decisionmaking and facilitates cognitive function that is important for improved quality of life.


INTRODUCTION
To perceive color from our surroundings is an innate ability that becomes perfected with daily experiences. The human brain needs a complex visual network to achieve this color processing [1,2,3]. While it is true that color perception varies across different cultures [4,5], the use of colors is spreading in various fields. In medicine, different types of colors have been used to assess cognitive function for diagnostic purposes [6,7] [8,9,10]. The Ishihara chart can detect impairment of color perception [11,12]. Color has also been used therapeutically in cancer patients [13,14]. However, patients come from different educational backgrounds, and the results of neuropsychology tests might vary depending on background educational levels. This could have implications for management strategies. It is therefore important to investigate the influence of education on color processing. Education develops our skills, knowledge, problem-solving abilities, and other cognitive functions [15,16] by improving the brain's ability to adapt to new stimuli and challenges [17,18]. Higher education is associated with higher cognitive functioning [19], but this can decline in elderly populations [20]. Influence of education on theta [21] and delta [22] rhythms, as well as its influence on processing of different shapes and their arrangements [19,23], has been investigated previously. However, no study has investigated the influence of education on perception of different colors. Therefore, here we investigated the influence of education on color processing using reaction time (RT) analysis and ERP tools.
RT measures the time interval between stimulus presentation and the subject's response [24]. Researchers have studied three types of RTs. Simple RT can be measured with one stimulus and one response; recognition RT presents a stimulus, but requires no motor response; and choice RT has multiple stimuli with multiple responses [25,26]. Jain et al. [27] showed that RT was faster in males and those who exercised regularly, compared with females and those with a more sedentary lifestyle. It is possible that faster RTs are underpinned by a faster nervous system processing time, corresponding to a faster muscular movement [28], which could be interpreted as meaning that a person with faster RT has higher cognitive function [29]. For these reasons, we used RT as a marker of cognitive processing speed in the present study.
In addition to RT, we also measured ERP components to assess cognitive function. Electroencephalography (EEG) is non-invasive, painless, and relatively inexpensive compared with other neuroimaging techniques [30,31]. In contrast to the excellent anatomical resolution of brain imaging techniques (3-4 mm for fMRI) that allow the visualization of the brain networks involved in a certain task, electrophysiological tools such as EEG, give fine temporal resolution (in the order of milliseconds), allowing us to observe the various ERP components representing the cognitive stages used during performance of a task. Thus, decreased amplitude and/or delayed latency compared with normal values can be interpreted as representative of a cognitive deficit. Such findings should be interpreted by taking into account the cerebral area from which the ERP is recorded [32]. All areas of cognition, such as perception, attention, selective attention, attention switching, resistance to distractive interference and memory, executive function linked to problem-solving, planning, and decision-making, are associated with ERP profiles. Cognitive ERPs allow for the assessment of the different stages used in the information processing stream during performance of a task. Every cognitive function is associated with various cognitive stages. Each of these stages is implemented within separate neural processes to achieve normal function [33]. Indirectly, non-invasive ERP components can therefore convey information pertaining to the functional connections between brain areas, and represents a tool with which to investigate cortical activity [34].
The ERP signal can directly measure the neural activity related to a specific event [35,36]. The recorded signal is mainly influenced by the electrical field of the head scalp, which comes from the current of the extracellular regions [37,38]. In response to visual stimuli, six ERP components have been identified: C1, P1, N1, P2, N2 and P3 [39]. Higher amplitudes of the N1 (N170) reflected increased attention in a study of different shapes and their organizations [19], and in a separate study of other race faces [40]. The N200 ERP component is a negative-going waveform that is evoked at 200-350 ms poststimulus, and reflects executive cognitive function [41]. The N200 has subcomponents of N2a, N2b, N2pc that mainly reflect activity in occipital-temporal regions. Higher N200 amplitudes are thought to reflect increased intention [42]. The P300 ERP component is a positive deflection, usually ranges from 250 ms to 900 ms, is typically evoked between 300-400 ms [43], and reflects activity at Cz and Pz areas [44]. The P300 has been identified as a marker for response inhibition [45], which involves the activation of the executive system of the frontal lobes [46]. The neural basis for this executive system is believed to be a distributed circuitry comprising prefrontal areas and anterior cingulate gyrus [47], the orbitofrontal cortex [48], the ventral frontal regions [49], and the parietal, dorsal, and ventral prefrontal regions [50]. Elevated amplitudes and shorter latencies of P300 components are related to higher attention [51] with increased consciousness [52]. P300 deficits have been reported in many disorders, but there is also an abundance of literature implicating other ERP components, such as mismatch negativity, sensory P50, and N400 (linked to language and semantic processes) [53,54]. Nevertheless, because both the N200 and the P300 ERP components are related to cognitive functions [23], we analyzed these two components to assess visual cognitive function. There is currently a lack of information regarding how education influences processing of different color stimuli in the human brain on an electrophysiological level. The main aim of this study was therefore to investigate the impact of education level on color processing (as an index of cognitive processing) using RT and ERP.

Study Population
The required sample size was calculated by a statistician using Power and Sample size (PS) software. Participants were recruited via e-mail advertisements, personal communications, and/or internet advertisements. Subjects were stratified according to their highest educational qualification, and accordingly grouped as low (G1, mean age±SD; 30.91±5.32 years, 8M and 9F), medium (G2, 29.89±8.54 years, 10M and 7F) or high (G3, 27.01±2.97 years, 8M and 9F) education. The educational groups were established according to the Malaysian educational system. The low educational group (mean education±SD; 8.93±2.63 years) had completed SPM (Sijil Pelajaran Malaysia), the medium educational group (12.79±0.80 years) had completed STP (Sijil Tinggi Pelajaran Malaysia) and the high educational group (15.88±1.36 years) had completed higher educational qualifications such as a diploma, Degree, Master's, or PhD.

ERP Net/Cap
A 128-electrode sensor net was used for data acquisition. This net was connected to a headbox or amplifier, and the net was elastic to stabilize the electrodes on the head. Each Ag/AgCl electrode was covered by small plastic pedestal with a sponge contained within. Prior to each participant using the net, we soaked it in an electrolytic mixture of KCl solution and baby shampoo. Spontaneous electrical activity can sometimes pass through the wet sponge of the net to the scalp, and we measured this activity in terms of ERP amplitudes and latencies on another computer located to another room.

RT and ERP Stimuli and Recording
RT and ERP recordings were performed in the MEG/ERP laboratory at Hospital Universiti Sains Malaysia (HUSM). The ERP recordings were made in a dimly lit, electrically shielded and sound-treated room. Subjects were comfortably seated 1 m in front of a 22" LCD computer. Different color stimuli were presented using Eprime software (v 2.0 software (Psychology Software Tools, Inc., Sharpsburg, Pennsylvania, USA). ERP recordings were made with a 128channel sensor net. 20 different colors were used as stimuli (13 × 17 cm): they were presented randomly on a white background for 1.4 s, with a 1.0 s inter-stimulus interval (ISI). Fig. 1 depicts the experimental paradigm used. Subjects were instructed to push 'button 1' for like and 'button 2' for dislike. All data were recorded on Net-Station software 5.2 (Electrical Geodesics, Inc., Eugene, OR, USA). Amplitudes and latencies of the N200 and the P300 ERP components were analyzed. For our purposes, we considered the observed RTs to represent "decision time". The experimental task used here differs from standard paradigms, in that it does not require the subject to make a correct/incorrect response.

Data Analysis
The band pass filter used was 0.3-30 Hz, with a stimulus rate of 0.5 Hz. Electrode impedance was within 0-50 KΩ. Using Net-Station software, we collected the mean differences between 'liked' and 'disliked' stimuli for the amplitudes and the latencies of the N200 and the P300 ERP components. Values from 19 channels were analyzed (FP1, FP2, F3, F4, F7, F8, Fz, C3, C4, Cz, P3, P4, Pz, T3, T4, T5, T6, O1, and O2 electrodes). Values falling within the -100-800 ms ranges were incorporated in the final analyses; the baseline was corrected to 100 ms before stimulus onset. Eye blinks, eye movements and movement artifacts were removed using the artifact detection tool in Net-Station software. Results were analyzed in Statistical Package for the Social Sciences 22 (SPSS22) software, using a one-way analysis of variance (ANOVA) analysis. The significance level was set as p≤0.05.

RESULTS
The grand average waveforms of the amplitudes and the latencies of the N200 and the P300 ERP components are shown in Fig. 2a (G1  'like' and 'dislike' color stimuli were collected for the amplitudes and the latencies of the N200 and the P300 ERP components (Tables 1 and 2).

Reaction Time (RT) Analysis
Our experimental paradigm did not allow for the assessment of correct vs incorrect responses; instead, participants chose whether they liked or disliked the different colors that they were presented with. Therefore, we were only able to analyze reaction times (RTs) an performance. RTs were compared across the three groups. A one-way ANOVA showed a main effect of group, F(2, 48) = 8.449, Between-group contrasts revealed that RTs in the higher education (G3) (mean±SD, 763.8±239.53 ms) group were significantly lower than those in the medium (G2) ( 907±242.50 ms) and low (G1) ( 1145.52±416.59 ms) education groups (Fig. 3). G2 showed reduced RTs compared with G1 ( 0.03).

The N200 Component
N200 amplitudes were clearly seen at seven electrode sites: T4, T5, T6, P3, P4, O1 and O2. Higher amplitudes were evoked in G3 compared with the G2 and the G1 groups across all channels, but these were only significantly different at the T6 site, F(2,48) = =0.036). Conversely, G2 had higher N200 'like' and 'dislike' color stimuli were collected for the amplitudes and the latencies of the N200 and the P300 ERP components (Tables 1 and 2).

Reaction Time (RT) Analysis
ntal paradigm did not allow for the assessment of correct vs incorrect responses; instead, participants chose whether they liked or disliked the different colors that they were presented with. Therefore, we were only able to analyze reaction times (RTs) and not task performance. RTs were compared across the way ANOVA showed a main (2, 48) = 8.449, p = 0.001. group contrasts revealed that RTs in the higher education (G3) (mean±SD, significantly lower than those in the medium (G2) (p = 0.03, 907±242.50 ms) and low (G1) (p = 0.0002, 1145.52±416.59 ms) education groups (Fig. 3). G2 showed reduced RTs compared with G1 (p = N200 amplitudes were clearly seen at seven electrode sites: T4, T5, T6, P3, P4, O1 and O2. Higher amplitudes were evoked in G3 compared with the G2 and the G1 groups across all channels, but these were only significantly (2,48) = 4.497, p =0.036). Conversely, G2 had higher N200 amplitudes at 6 of the 7 sites (T4, T6, P3, P4, O1 and O2) compared with G1 (Table 1). In terms of the N200 latencies, G3 possessed the shortest latencies of the three groups at most of the locations (5 out of 7 locations: T4, T6, P4, O1 and O2). The next shortest latencies were in G2 (5 out of 7 locations: T4, T5, T6, P4 and O2; see Table 1).

The P300 Component
The P300 results were consistent with those observed for the N200 component. Again, G3 expressed the highest P300 amplitudes at most of the sites (6 out of 7 locations: T4, T5, T6, P4, O1 and O2) compared with the G2 and G1 groups ( Table 2). G2 had the next highest amplitudes at most of the locations (5 out of 7: T4, T6, P4 and O2) compared significant group effect was observed at T6, F(2,48) =3.186, p =0.047 (Table 2). There was no consistent relationship between group membership and latencies of the P300 component (Table 2).

DISCUSSION
In this study, we investigated the influence of education on color processing (indexed by RTs and ERPs) in low (G1), medium (G2) and high (G3) educational groups. G3 displayed significantly shorter (i.e. faster) RTs than the other groups, and G2 had faster amplitudes at 6 of the 7 sites (T4, T6, P3, P4, O1 and O2) compared with G1 (Table 1). In terms of the N200 latencies, G3 possessed the shortest latencies of the three groups at most of the of 7 locations: T4, T6, P4, O1 and O2). The next shortest latencies were in G2 (5 out of 7 locations: T4, T5, T6, P4 and O2; see The P300 results were consistent with those observed for the N200 component. Again, G3 expressed the highest P300 amplitudes at most of the sites (6 out of 7 locations: T4, T5, T6, P4, O1 and O2) compared with the G2 and G1 groups ( Table 2). G2 had the next highest amplitudes at most of the locations (5 out of 7: T4, T6, P4 and O2) compared with G1. A significant group effect was observed at T6, =0.047 (Table 2). There was no consistent relationship between group membership and latencies of the P300 In this study, we investigated the influence of education on color processing (indexed by RTs and ERPs) in low (G1), medium (G2) and high (G3) educational groups. G3 displayed significantly shorter (i.e. faster) RTs than the other groups, and G2 had faster RTs than

Fig. 2c. Grand average waveforms positions in the high education group (G3). Blue represents 'like' and red represents 'dislike' in relation to the colors displayed
G1.The highest amplitudes and shortest latencies of the N200 and P300 ERP components were evoked mostly in the G3 and G2 groups. The T6 area showed a significant group effect for amplitudes of 6

Fig. 2c. Grand average waveforms of the N200 and P300 ERP components at 19 electrode positions in the high education group (G3). Blue represents 'like' and red represents 'dislike' in relation to the colors displayed
G1.The highest amplitudes and shortest the N200 and P300 ERP components were evoked mostly in the G3 and G2 groups. The T6 area showed a significant group effect for amplitudes of both the N200 and the P300 components (Tables  1 and 2). Subjects in G1 (low education level) displayed the shortest amplitudes and the longest latencies for both components across most of the sites.

Fig. 2b. Grand average waveforms of the N200 and P300 ERP components at 19 electrode positions in the medium education group (G2). Blue represents 'like' and red represents of the N200 and P300 ERP components at 19 electrode positions in the high education group (G3). Blue represents 'like' and red represents 'dislike'
both the N200 and the P300 components (Tables s in G1 (low education level) displayed the shortest amplitudes and the longest latencies for both components across Participants in G3 had significantly faster RTs than those in G2 and G1. G2 participants had longer RTs than those in G3, and G1 had the longest RTs of all groups. There was a significant main effect of group on RT, as depicted in Fig. 3. Faster RTs are directly associated with faster decision-making [55]. As a corollary of this, we suggest that higher education is associated with faster decision-making.
Being an early ERP component, the N200 reveals basic information about 'bottom-up' sensory processing of stimuli. The P300, however, reflects 'top-down' perceptual and cognitive processing of stimuli [56]. Higher N200 amplitudes are associated with higher attention [57], and P300 amplitudes are directly proportional to successful perceptual and cognitive processing [58]. Kok et al. (2004) found that higher amplitudes of the P300 component reflect higher levels of cognitive function, including, but not limited to orientation of attention, response modulation, and response resolution [59]. Increased attention results in  quicker decision-making [55] and makes use of higher cognitive functions [29]. Taken together, these findings suggest that education improves attention, cognitive processing and improves cognitive functioning in general, which is associated with faster decision-making. Higher amplitudes for both components were evoked in the high education group (G3). However, we note that shorter amplitudes were recorded at some locations in G3, and our results should be interpreted with caution. On the other hand the low education group took the opposite interpretation of the high education group.
The N200 latency has an expression on the basis of the discrimination and classification of the visual stimuli that means on the basis of the rapid registration of the visual input which has positive correlation with the latencies of the N200 [60]. Applied to the present study, this suggests that participants in the high education group were able to rapidly register the color stimuli they were presented with: G3 evoked the shortest latencies at most of the electrode positions compared with G2 and G1. Indeed, G1 had the longest latencies, suggesting much slower registration in this group.

Fig. 3. Reaction times (ms) across high, medium and low educational groups
The P300 latency reflects the time taken to attend to and process a stimulus [61], relying on cognitive processes comprising attention, learning, and decision-making. The fact that there were (albeit non-significantly) longer latencies of the P300 component at most of the electrode locations in the lower education group could reflect longer processing times underpinned by attention, learning and decisionmaking processes in this group. The higher education might made the G3 and then G2 groups in more quicker timing process of cognitive functions with the evidence of the significant and non-significant shorter latencies of the P300 component (Table 2). However, higher and lower amplitudes, and longer and shorter latencies were expressed in a mixed manner across all three groups, making it difficult to find a parsimonious interpretation of the results.

CONCLUSION
We investigated the influence of education (indexed as low, medium or high) on color processing (indexed by RTs and ERPs). Results suggest that participants in the high education had higher attention and cognitive functioning, as well as improved learning and decision-making. Additionally, the high education group was able to process stimuli faster than those in the other groups, presumably because of their increased attention. To conclude, higher education is associated with increased attention, which itself is related to faster decision-making. This ultimately leads to higher cognitive functioning and improved quality of life.

LIMITATIONS
ERP studies are associated with certain limitations with poor spatial resolution and interindividually there are big differences of amplitudes and latencies of the ERP components [32]. Though electrophysiological components analysis alone can give us information about heterogeneity among the groups, but together with endophenotype analysis could reduce the impact of heterogeneity [62,63]. Another limitation of our study was small sample size. Larger sample of participants might be more reliable to generalize the results to a larger population.