Empirical research of language and Categorical Perception of colour

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

3. Empirical research of language and Categorical Perception of colour

Behavioral, neuroimaging and electrophysiological studies have used the categorical perception (CP) of color to demonstrate the Whorfian Hypothesis. The categorical perception refers to how we categorize things in the world (objects, people ideas, events etc.) and it is believed that language determines CP in adults. The best known method in testing the CP is color categories. (Harnad, 1987) Although, the spectrum of color is continuous, and all humans with normal vision see the color in the same way, languages of the world divide the colors into different categories using linguistic terms present in our language (e.g. blue, green, red). Speakers of different language organize colors in different ways, and the presence of categorical borders between colors accelerates the discrimination. What drives the discrimination of colors is not the perceptual cue of the human vision, but rather the categorical information that is embedded in the language. This is evident in how we respond to the color, for example, it is easier to distinguish between two colors that are from different category (between category), than two colors that are from the same category (within category). The color CP effect establishes a link between both color perception and color categories, and the most accepted account of CP is the naming strategy, basically research on the lexical encoding of color. (M.H. Bornstein and N.O. Korda, 1984)

Behavioral studies have measured the reaction time (RT) and the accuracy (AC) to see whether different linguistic terms make a difference in color discrimination. Neuroimaging studies have looked directly into the brain to investigate whether language areas are active in color perception. Electrophysiological studies have explored the time course of the brain response in color detection.

The following experiments investigate the origin and the nature of color categorization concerning how Categorical perception of color (Harnad, 1987), linguistic relativity (Whorf, 1956) and cognition are related.

3.1.1 Behavioural studies

Winawer et al. (2007) tested English and Russian speakers that categories color differently. Unlike the English speakers, Russian speakers are always able to distinguish between two different blues i.e. dark blue (siniy) and light blue (goluboy). For the English speakers these shades of blue felt into the same basic color category, namely, blue. The aim of this study was to test whether these linguistic differences in the two language groups have an effect on color discrimination. The two groups of subjects have to perform a simple visual searching task, in which Subjects were presented with three color squares in a triad form of twenty different shades of blue (fig. 2), and were asked to say which one of the two bottom squares is identical to the top one.

Fig. 2: An example of the triad color task.

The distractors target were either from the same Russian category as the reference (e.g. within-category - both were goluby or both were siniy), or from different Russian categories as the reference (e.g. between categories - one was goluby and one was siniy). Three experimental conditions were setup: No interference, verbal interference and spatial interference. In the no interference condition there was no use of language and memory involvement, in the verbal condition the subject had to silently repeat eight digit numbers in mind, while doing the color task, and in the spatial condition the subjects were instructed to maintain a grid pattern in mind until tested. The stimuli were presented until the subjects pressed a button. The subjects at the end of the experiment had to do a color classification task to define the color boundaries of each subject. The behavioural data were analysed relative to their own boundary.

Fig. 3: Reaction Time for all the conditions. Between- and within-category comparisons are included in these graphs.

The results showed (see fig. 3) that Russian speakers were faster in discriminating color between-category (e.g. one goluby and one siniy) than within-category (e.g. both goluby or both siniy). This category effect for the Russians disappeared by verbal interference, but in the spatial interference, demonstrating the effect of language is online. English speakers did not show in any condition a category effect. This communicative necessity of the Russian speakers, as the results show, appears even when performing a visual task that does not require language. These demonstrate that language categories are involved in perceptual decision and differ across language groups. (Winawer et al., 2007)

If language is really involved in perceptual decision, then the right visual field, which is known to be related to the left cerebral hemisphere (LH), the dominant hemisphere for language in most adults, should be more involved. Gilbert et al. (2006) found evidence about this proposal in a behavioural color task. English subjects were presented with a visual search task in a ring of 12 colored squares. The colors stimuli[1] were two shade of blue and two shade of green (fig. 4). The Participants were required to press a button (left or right index finger) of a single target color, indicating the side in which the target appeared in the ring, among 11 identical distractors. The visual task remained displayed until a decision was made.

Fig. 4: (a) Print-rendered versions of the four colors used. (b) Sample display for the visual search task.

The target and the distractors were different in hue and were either from within category of the distractors (e.g. One blue among another blue) or from between category of the distractors (e.g. blue among greens). The results show that subjects detected the target faster in between- than in within-category with a RVF advantage (fig. 5-c).

Fig. 5: (c) RT of the no-interference condition of the RVF and the LVF between within- and between-category. (d) RT of the verbal interference.

This confirms that the Whorf hypothesis is supported more strongly in the RVF, which reflect the LH, the dominant part of the language.

To support this interpretation Gilbert et al. (2006) have run a similar second experiment. The participants were presented the same visual search task, but this time with a verbal interference task, in which the subjects had to keep in mind eight-digit number, while doing the experiment. The idea is that the RVF advantage in discriminating between category (experiment 1) is by reason of the activation of linguistic information, with a verbal interference condition the RVF effect should disappear. As expected the results of the second experiment show in the verbal condition the RVF effect disappeared (fig. 5-d), confirming the idea of the lateralized Whorf hypothesis, i.e. color perception is more lateralized to the LH. In other words, it seems that half of the perception (RVF) of the reality is filtered by the native language and the other half is not (LVF). (Gilbert, Regier, Kay, & Ivry, 2006)

Further evidence of the lateralized hypothesis were achieved by Drivonikou et al. (2007). The Authors conducted two new experiments to confirm and support the results of Gilber et al. (2006). They reanalysed a color search data from Daoutis et al. (2006) and used two different approaches to confirm the lateralized Whorfian effect. In the Daoutis et al. experiment the subjects were asked to identify if the target was present in a display of multiple stimuli, i.e. 4, 16 or 36 at a time. The target was present in half of the trials and could appear across 32-target locations. The stimuli were presented randomly, selected from blue, purple, pink and green. A second experiment was done with different stimuli set, blue, green and purple. Daoutis et al. experiment shows that the reaction time of the subjects to detect the target was influenced by the categorical relationship between target and distractor colors. In others words, the participants were faster in detecting between-category targets than within-category ones in a crowed display. (Daoutis, Pilling, & Davies, 2006)

In the reanalysis of Drivonikou et al. (2007) the target could appear on the left or on the right of the central fixation, where the trials appeared, instead of across the 32 target locations in Daoutis et al. (2006) experiment. The targets were either from between- or within-category. The two experiments differed only for the color target: Blue and Green in the first experiment and blues and purples in the second experiment. The results showed a faster response in between-category than in within-category. The reaction time was slower in comparison of the Daoutis et al. results, but with lager category effects. A category effect in the RVF was also found, i.e. a faster response was noted when the target was presented on the right side of the screen. A weaker effect was, however, found in the LVF (RVF was 60ms faster than the LVF). This evidence showed a similar effect to that of the Gilber et al. (2006) results, a stronger activation of the RVF than the LVF with two different visual tasks. This confirms the claim of the lateralized Whorfian effect. (Drivonikou & Kay, 2007)

Further studies demonstrate that the lateralized category effect could be entirely the result of new learned categories. (Zhou et al., 2010 and see also Drivonikou et al., 2011). Franklin et al. (2008) analysed the color CP in infants. The data evidenced that infants aged 4 to 6 months old are also quicker at discriminating between colors from between categories than colors from within category. But in adults it seems that the color CP is lateralized in the left hemisphere (LH) due to the influence of the linguistic terms, whereas in infants it seems that it is lateralized in the right hemisphere (RH), due to the fact that the color terms have not yet been learned. What they showed is that the switch from the RH to LH in color CP occurs when the words of the different categorical color border are learned. (A. Franklin, Drivonikou, Bevis, et al., 2008; A. Franklin, Drivonikou, Clifford, et al., 2008; A. Franklin, Pilling, & Davies, 2005)

While all these studies already presented in this chapter support the lateralized color effect theory, other studies have found only a category effect (Liu, Chen, Wang, Zhou & Sun, 2008 and Winaver et al., 2007) and some others did not even find color discrimination effect in perceptual decision. (Brown et al., 2009; Lindsey & Brown, 2009; Pinto, Kay & Webster, 2010)

3.1.2 Neuroimaging studies

Despite growing behavioural evidences on the effect of language on the categorical perception of color, several neuroimaging studies were motivated from the work of Gilbert et al. and Drivonikou et al. to analyse the neural network activity of the Whorfian effect. In fact the neuroimaging techniques give the opportunity to look directly into the brain and analyse whether language components alter our cognitive processes.

Tan et al. (2008) measured with the functional magnetic resonance imaging (FMRI) the brain activity to determine, whether the language in color CP modulates the neural network. In a typical color visual search task, the participants were presented simultaneously to two colored squares of 100ms with a grey background and followed by a mask of 900ms. Two conditions were setup for the experiment. The “easy to name” condition and “hard to name” condition and the participants had to judge whether the two viewed colors were the same or different. See stimuli setup below (Fig. 6).

Fig. 6: A: are easy-to-name colors. B: hard-to-name colors.

The authors tested the accessibility of the color stimuli in a premilinary behavioural experiment to subjects who did not took part in the formal experiment. The subject’s task was to say aloud the name of the colors within one second. All the subjects respond within the time in the easy to name stimuli but not to the hard to name stimuli.

For both conditions, different cortical regions were activated: the visual cortex, critical for color vision, and the frontal gyrus in both left and right hemisphere, critical for inhibition and intentional control. The language–related regions were activated only in the easy to name condition, i.e. left superior temporal gyrus and parietal regions, two areas that are involved in word finding processes. These results suggest that language is automatically involved in color perception, as soon as the colors compared are easy to be named. (Tan et al., 2008)

A slightly modified version paradigm of Gilbert et al. was used by Siok et al. (2009) in a FMRI study to analyse the neural correlates activity of the brain in a visual search task. In contrast to the Tan et al. experiment, this study was designated to look into neural substrates of the behavioural RVF advantages and to clarify why linguistic information aids regions of color vision in the brain. The participants were asked to identify if the target color appears on the left or on the right side of a colored ring[2].

The findings suggest a stronger activation when the target and the distractors where form different lexical category, in the RVF than in LVF of the language-related areas involved for color CP, i.e. the posterior temporoparetial region, the middle temporal gyrus, and the inferior prefrontal cortex. Also activation of the visual area V2/3 was found and the activity of the V2/3 increased when the colors from different lexical categories were presented in the RVF. Any brain language regions were not associated with a stronger activation in the LVF (fig. 7).

Schermata 2014-01-06 alle 19.22.25.png

Fig. 7: Brain regions with significant activation during the recognition of colors in comparison of between- and within-category in the right visual field.

The Authors suggest that the color CP provokes a neural activity including the posterior temporoparietal and V2/3 regions that determine the Whorfian effect. Surprisingly, both Siok et al. and Liu et al. 2009 could not replicate the results of Gilbert at al. in the behavioural level. Similar results of Siok et al. in the fMRI level were found in the study of Ikeda and Osaka (2007) that observed a left lateralization of color perception in the brain. (Siok et al., 2009 & Ikeda and Osaka 2007)

Kwok et al. (2011) analysed the neural network of the Whorfian effect before and after the acquisition of new named categories. In this study, they wanted to show that acquiring artificially new names for the universal color names green and blue for 2h increase the volume and the density of gray matter in the virtual cortex V2/3. Past researchers have already shown a structural plasticity in the acquisition of new skills, but these were obtained in a training period of a number of weeks or even years, not in a short period. The subjects were exposed to intensive training to map new terms for color categories, i.e. two new terms for green and two new terms for blue. The duration of the training was three days and involved five sessions in three activities: listening, naming and matching. Kwok and colleges collected high-resolution 3D images of the magnetic resonance imaging (MRI) of the participants before and after the training in a visual fast-mapping task[3].

The results show, comparing both pre- and post-training brain scan, that there is a significant change in the visual cortex V2/3 and in the cerebellum. In other words, the grey matter increase rapidly in these regions, especially in the left visual cortex. These results not only confirm the recent publications of the lateralized Whorfian effect, but also that the human brain is structurally more plastic than previously believed. (Kwok et al., 2011)

Several behavioural studies have evidenced that language facilitates color discrimination by measuring reaction time and accuracy in visual search tasks. (i.e. Drivonikou & Kay, 2007; Gilbert et al., 2006; Winawer et al., 2007) Neuroimaging studies have shown that language is automatically involved in color CP and has been found in language related brain regions, as well as in the virtual cortex. (Kwok et al., 2011; Tan et al., 2008; Siok et al., 2009 & Ikeda and Osaka 2007) However, how the brain reacts and responds in color discrimination has been investigated using the event related potential (ERP) technique. (Luck, 2005) Using this technique researchers have recoded the electrical activity of the brain to explore the time course and the mechanisms that are involved in color perception.

3.1.3 Electrophysiological studies

The first studies in color CP using the electrophysiological technique was done by Fonteneau and Davidoff (2007). The authors used a passive oddball paradigm[4] to investigate whether different color categories effect a change in the temporal dynamics of neural responses. In this study, color was not the relevant task, and a judgement of the colors was not required. The participant’s task was to detect infrequent cartoon characters among blocks of color patches. The experiment involved six different blocks and each block was presented two different color stimuli, a frequent color (standard) and unfrequented color (deviant). The colors were either from the same category, within category (two different greens) or between category (one blue and one red) and they differed only in hue. The stimuli were presented for 200ms with a interstimulus of 600 ms. The results show that deviants stimuli elicited a greater ERP amplitude waves form for both between- and within category in a time range of 160-200ms and extended to the 160 -280 ms only for within category. The ERP waves showed differences between the between- and the within-category comparison. Respectively, the peak latency of the between category occurred at 195ms, 19ms earlier than for the within-category. In the visual ERP sequence[5] the components found are related to a late N1 extending to N2 component (correspond to a post-perceptual stages) that does not show clearly, if color CP can occur before the completion of the perceptual processing. The authors interpreted these results as the first evidence for neural correlates of color differences that are used during passive visual events and also a more difficult discrimination for within- than between pair. This experiment did not found any lateralization, and the authors argue that the perception of color is not due to a lateralized neural substrate, but it reflects an on-line top-down influence of color labels from language areas in the cortex. (Fonteneau & Davidoff, 2007)

Holmes et al. (2009) carried out an experiment similar to Fonteneau and Davidoff’s, but the experiment setup differed in a number of ways. First, each block contained one standard stimulus and two deviant stimuli, one from the between-category of the standard (two different blues) and one from a within-category (one blue and one green). In the Fonteneau and Davidoff´s experiment just one standard and one deviant were used. Secondly, they used smaller hue differences for within-pair colors, defined by the author as more typical setups used normally in color CP experiments and each stimulus was presented for 400 ms in the middle of the monitor on a grey background. They recorded the brain potential in a color detection task across the blue and green boundary to analyse the time course, and the differences for between- and within-category.

The ERP waves showed shorter latencies to between-category in contrast to within-category deviants and P1 (80-120 ms) and N1 (130-190 ms) ERP components were found that reflects an “early phase” for perceptual processes. In contrast to Fonteneau and Davidoff´s results, the color CP modulation was not clearly interpretable as an early perceptual stage, since the peak was a late N1 component. Also, a late effect of P2 and P3 components were found in between- than in within-category, demonstrating the influence of color categorization on post-perceptual process as well. They also measured the CP effects using behavioural experiments. Reaction time (RT) and accuracy (AC) data demonstrated that between-category in color discrimination were faster and more accurate than of within-category stimuli. The results of the study show the influence of color categories at early perceptual stage with a category effect around 90ms. (Homes, 2009)

Thierry et al. (2009) found similar evidence using brain potentials. The aim of the study was to investigate whether language affects the perception of color down to the level of the visual process or if it is just a post-perceptual cognitive process. Using the ERP technique, they measured the brain response of Greek and English speakers in a visual color oddball task (fig. 8) to demonstrate that the existence of two color terms in Greek, ghalazio and ble, for two different shapes of blue leads to a faster response in color discrimination than for English speakers, which only one linguistic term “blue” exists. The participants were presented with four experimental conditions: in two conditions the stimuli were dark or light blue and in the other two conditions the stimuli were dark or light green. The subjects were instructed to press a button only when the repeatedly stream of circles (standards) change in a square shape (target).

Fig. 8: Experimental design and sample of stimulus sequences presented in the four experimental blocks.

In each block, 70% of the standard stimuli were dark or light colors, 10% of the deviant stimuli were contrasting luminance colors of the standard (e.g. dark colors if the standards were light and vice versa) and the remaining 20% were the targets. The green and blue stimuli were setup in terms of saturation and luminance and presented for 200ms with an interval of 800ms. The visual color oddball paradigm was designed to produce a deviance component known as the Visual Mismatch Negativity (vMMN), a component of the ERP elicited by perceiving infrequent deviant stimuli and usually considered automatic and preattentive. As expected, they found a greater vMMN for Greek than for English participants with mean signal amplitude between 100-130ms (P1 peak), by reason of the existence of two basic shades of blue. They found a triple interaction of color, participant group and deviancy, but ambiguously not a color effect. Hovewer the MMN was greater fro blue than green deviant stimuli for Greek speakers. In contrast, English participants showed a similar magnitude for blue and green conditions. These electrophysiological findings confirm that the perception of colors is preattentive and unconscious related to the native language.

The previous studies have shown a preattentive categorical perception of color, but to support the finding of fMRI and behavioural studies for the lateralized Whorfian effects, Mo et al (2011) used a different oddball paradigm to test if the left lateralized effect of language in color CP occurs also in “early phase” of perceptual process. Participants performed an oddball task (fig. 9) while brain activities were recorded. In each trial the two color squares were presented for 200 ms with a fixation mark in the middle of the screen. The participant´s task was to focus on the fixation mark, a plus sign (“+”) and press a button when it turns in a circle (“o”). No instruction was given about the colors.

Fig. 9: (A) Print-rendered version of the four colors used. (B) Experimental design and a sample of stimuli presented in the four experimental blocks.

Four colors equivalent in hue and luminance were used: two different shapes of green and two different shapes of blue. In each block the standard stimuli were one of the adjacent color (green or blue) and the deviant pairs were one of the stimuli colors (G1-G2 or B1-B2). The deviant stimuli were distributed evenly to the left and right positions and were from between- or within-category. Analysing the ERPs waves showed greater amplitude of the vMMN to between-category than within-category deviants when the deviant were presented in RVF, while the effect was absent in the LVF. The deviant amplitude peak began at 130 ms post-stimulus ended at 190 ms and peaked at 160 ms, showing a N1 component. (see fig. 10)

Fig. 10: A and B show ERPs and maps of standard and RVF deviants. C and D show ERPs and maps of standard and LVF deviants.

The authors found a significant interaction in the vMMN for visual field and colors pair in the RVF, presuming that the LH, the dominant hemisphere for language in most adults, controls the color perception. The electrophysiological result evidenced that color discrimination occurs in the early stages of perceptual processing and that color difference is lateralized to the RVF.

3.4 Conclusive remarks and other evidence supporting the Whorfian effect

The various neuroscience studies in color CP have shown that language plays a role in color perception, i.e. faster discrimination in between- than within-category, activation of language regions in the brain and occurs in an early stage of perceptual decision. The majority of the experiments have confirmed the Whorfian effect (Athanasopoulos, 2009; Fonteneau & Davidoff, 2007; Holmes, Franklin, Clifford, & Davies, 2009; Liu et al., 2009; Tan et al., 2008; Thierry, Athanasopoulos, Wiggett, Dering, & Kuipers, 2009; Winawer et al., 2007;) and some others have found that the Whorfian effect is lateralized in the LH. (Drivonikou & Kay, 2007; Gilbert et al., 2006; Mo, Xu, Kay, & Tan, 2011; Siok et al., 2009; Ikeda & Osaka 2007). Several studies, however, could not find a lateralized effect (Fonteneau & Davidoff, 2007; Holmes, Franklin, Clifford, & Davies, 2009) and some others could not replicate the findings of Gilbert et al. (2006) in the behavioural level (Liu et al., 2009; Siok et al., 2009). The lateralized hypothesis is still widely debated and represents an important effect that remains to be confirmed through further research.

In fact, to archive the lateralized hypothesis Witzel and Gegenfurtner (2011) re-implemented six different versions of the Gilbert et al.´s experiment and four versions of the Drivonikou et al.´s experiment. They paid attention to the color setup (used RGB and Munsell values) and eye movement of the subjects, because one hypothesis is that the RVF effect appears when participants maintain eye fixation to the center of the screen (see experiment of Roberson and Pak, 2009). Moreover, they repeated the experiment also with non-German student to exclude the doubt that the lateral effect does not appear because of the particularities of the German color categories.

Schermata 2014-01-06 alle 20.18.18.pngSchermata 2014-01-06 alle 20.16.11.png

Fig. 11: Results of the six re-implemented experiment of Gilbert et al. 2006. Panel on the left side show the reaction time for the main condition and in the right side represent the verbal condition. The row (a, b) is the original result of Gilbert et al. The rest are average of RT of all the six re-implemented experiments of Gilbert et al.

The data of RT and AC (see fig. 11) does not show any lateralized effect in all the six experiment, as the original hypothesis, instead they found the typical category effect in both visual fields. These also when the participants maintained a central eye fixation on the screen and in the version of non-Germans they even found a LVF effect than a RVF advantage. Also none of the four replications of Drivonikou et al.´s experiment showed a lateralized category effect, but also just a category effect. The Authors concluded that the ten experiments support and confirm the existence of a general category effect, and does not show any lateralization effect reported from some studies. This shows that the lateralized category effect is still a highly questionable hypothesis and further studies should be performed to find new evidence to understand the factors that elicit the literalized category effect.

Some critics could be done in the methodology used to setup some of the experiment described above. First of all, are the boundaries between green and blue a testable experiment for the Whorfian effect? In one side, is true that several languages in the world do not make this difference, since they have just one lexical term, but the discrimination is still easy since the two colors are different. Is not the same in testing light blue and dark blue, in which they have similar proprieties (i.e. difference just in luminance), and subjects have to do a more fined distinction between the two shades of blue than for blue and green colors. What might play also a role in the discrimination of blue and green colors is that the natural environment may account for a greater sensitivity for differences in luminance for green colors than for blue colors or in the other way around for example. A second critic, in my opinion, is that much of the experiments did not use a tachistoscopic presentation of the stimuli, on the contrary the stimuli were presented until the subject responded to the colors. (Drivonikou et al. 2007; Franklin et al. 2008; Gilbert et al. 2006; Winaver et al. 2007) This make the discrimination easier for the participants, instead it should be more difficult to be surer that nothing else plays a role (e.g. memory, language, through etc.) in responding to the colors, e.g. presenting the stimuli just for 100 ms. To sum up the recent empirical research described above confirms that the language plays a role in the perception of colors, the lateralized effect is still an open question, since just some experiment found the effect and the experimental setup of some study could be improved to be certain of the evidence found.

The Whorfian Hypothesis was investigated not only in color perception, but also in other fields of research, in which the language may affect other aspects of non-linguistic cognition. The main experiments include time perception: different cultures have different mental representation of time. For example, English speakers tend to use horizontal metaphors “front” and “back”. Mandarin speakers, on the other hand, tend to use vertical metaphors “up” and “down” for the perception of time. (L Boroditsky, 2001; Lera Boroditsky, Fuhrman, & McCormick, 2011; Chen, 2007; January & Kako, 2007) Spatial relation is another domain in which the Whorfian hypothesis is being investigated. It refers to the fact that different cultures organize, acquire and revise the spatial environment differently. (Lera Boroditsky et al., 2011; Bowerman & Choi, 2003; Choi & Bowerman, 1991; Hespos & Spelke, 2004; Levinson, Kita, Haun, & Rasch, 2002; Levinson, 1996; Li & Gleitman, 2002; McDonough, Choi, & Mandler, 2003; Niraula, 2004) In the Numerical cognition experiment is to test the different naming systems of different languages, for example, French speakers pronounce the word “thirty” before the word “five” when naming the number 35, Dutch speakers, in contrast, do it the other way around. (Casasanto, 2005, 2008; Gordon, 2004)

Different types of experiments have been done, which are supportive of the Whorfian hypothesis. The different experimental evidence, non color-related-paradigms beyond the color CP studies, contributes significantly to the topic. Although, all the findings described above have interpreted the data as the Whorfian effect; nevertheless researchers have not clear addressed the relation between them. What actually happen in the brain when the Whorfian effect is present? How such mechanisms are generated in the brain? These are questions that must be answered to really understand why different languages organize the perception differently. The key to these questions is to comprehend, how the connection of different cognitive systems that are located in specific places in the brain occurs, and how the internal structure and the interaction works. In the next section I will focus on a possible neurofunctional explanation of the whorfian effects that is given by the cognitive theory that has put the basis to explain human behavior by understanding the thought process.

[1] English speakers mark the distinction between blue and green, but the majority languages in the word have just one single term for those colors.

[2] See Gilbert at al (2006)

[3] See Markson et al. (1997)

[4] The oddball paradigm is a method used in evoked potential research, in which trains of stimuli that are usually auditory or visual are used to assess the neural reactions to unpredictable but recognizable events. (Luck, 2005)

[5] P1, N1, P2, N2 and P3 are ERP sequence elicited by an oddball paradigm, where P and N are the positive and the negative-going components. The P1 and N1 components correspond to the early phase of the perceptual process in the brain. The P2, N2 and P3 components correspond primarily to the post phase of the perceptual process. (Luck, 2005)