HEMISPHERIC DIFFERENCES AND PRIORITYBETWEEN LOCAL AND GLOBAL PROCESSING (ABSTRACT)Our research aims to work outwhether global precedence plays a major role in visual perception. It wouldalso indicate whether this means that the right hemisphere can process thelarger picture faster than the left hemisphere can process the finer details.
Theexperiment that we are conducting is a replication of a part of David Navon’sglobal precedence investigation mentioned in his paper, (Navon, 1977).In theexperiment, the subjects had to respond to a visual stimulus in the form of aletter on a screen. Their reaction time was recorded and results were formedfrom this. The results showed that the right hemisphere was slightly faster atprocessing the global information which was represented by the faster reactiontime when both the local and global detail was corresponding. The findings havebeen similar to the research of Navon who proved that the right hemisphereprocesses global information faster than the local information. (INTRODUCTION)Perception is commonly thought of as ability to interpretinformation that sensory organs transmit as well as using prior knowledge toform the basis of understanding. This also includes how a subject would respondto the information received from the environment. The process of perceiving isvery complex and uses multiple specialised brain structures in order to form areaction to a certain stimulus.
A major problem for psychologists is toexplicate how sensory inputs are somehow converted into perceptions of realobjects to “form the basis of perceptual experiences”, (McLeod, 2008).Visual information is processed through the visual cortexpart of the brain, near the back end of the brain. The cerebral cortex is thelargest part of the brain and plays a key role in awareness, language, memory,attention and many other processes, (Kandel, Schwartz, Jessell, 2000). When visual information is processed, thestimulus enters through the optic nerve and the signal is sent down theoccipital lobe, located at the back of the cerebral cortex. Just before thesignal is received by the occipital lobe, the signal is taken through thelateral geniculate nucleus which separates the data into a stream that containscolours and fine details. 1It is often assumed that the humanbrain processes the global and local properties of visual stimuli in alocalised fashion, with the left hemisphere specialising in local detail, andthe right hemisphere specialising in global formations.
Visual processing has been divided into two differentcategories: top-down and bottom-up. It would make sense to combine boththeories together to form the basis of a conventional theory, however, over theyears, theorists have chosen to side with one or the other with someemphasising on top-down processing and the others focusing on bottom-upprocessing. Gregory’s constructivists theory stated that perception is a”constructive process which relies on top-down processing”, (Gregory, 1970).Information that we acquire from the environment is often too ambiguous for usto understand and comprehend, so we require higher cognitive knowledge orexperiences to make inferences about what we perceive. Top-down is thereference to the use of the contextual information to understand what weperceive. An example of this theory is the fact that we can understanddifficult handwriting within a sentence better than isolated words.
2In contrast to Gregory’s concept, Gibson introduced the ‘directperception’ approach. (Gibson,1972) His theory suggested that environmental data is all that isnecessary for visual perception and that there is no need for any other piecesof information to perceive our surroundings. According to McLeod, Gibson’sbottom-up system implies that perception “involves innate mechanisms forged byevolution and that no learning is required” (McLeod, 2008). (EXISTING RESEARCH)David Navon suggested that we perceive global features andthen progress towards a more detailed analysis of local features. He alsoexplored the relationship between local and global perception which ismentioned in his report, (Navon,1977).
Navon’s experiment’saim was to test global precedence in perception by measuring the reaction timeof the global processing and comparing it to the reaction time of the localprocessing. He did this by asking subjects to respond to images that consistedof larger letters made up of smaller letters. Sometimes the smaller letterswould be correspondent to the larger letters and sometimes they would not.
A great number of different studies have been conducted, tofurther improve our understanding of the complex procedure of visualprocessing. An example of this is Rebecca Chamberlain’s research into artistsand observational drawings. Individuals with enhanced talents in visual arthave been shown to display heightened local visual processing skills. The aimof Chamberlain’s study was to “assess whether local processing biasesassociated with drawing ability result from a reduced ability to cohere localstimuli into global forms, or an increased ability to disregard global aspectsof an image”, (Chamberlain,McManus, Riley, Rankin, Brunswick, 2013). This helps reveals the brainsability to prioritise the certain parts for local and global processing, andalso the reason as to why it would do so. It links into our investigation as italso displays why artists with better local processing skills and dominance inthe right hemisphere would produce better, detailed pieces of artwork.Another study, by Lamb andRobertson, investigates whether the reaction time is the correct variable tomeasure in order to understand the processing order and hemispheric dominanceof global and local information, (Lamb et al, 1989). Theexperiments conducted by Lamb supported the statements by Navon which suggeststhat the right hemisphere is dominant in global processing and the lefthemisphere dominant in the local processing.
However, it also considers theinterference of global distractors on the processing of the local information. This report by Lamb goes hand inhand with another study by Shihui Han. Han’s report questions the brain’sprocessing by examining “the neural mechanisms of functional asymmetry between hemispheresin the processing of global and local information”, (Han et al,2001).
Shihui Han used functional magnetic resonance imaging (fMRI) toview specific parts of the head and map the operational sections whenprocessing information. From this, he could confirm Navon’s study on thefunction of both sides of the brain when processing visual stimuli. The aim of the experiment weconducted was to test the ability of people to recognise smaller letters whichformed a larger product; which would show whether the brain prioritises globalor local processes. This experiment would also demonstrate the repeatability ofNavon’s Experiment.
(METHOD)(Participants)For the experiment, 9 healthystudents (5 males and 4 females, aged 16/17) volunteered to participate in theexperiment from The West Bridgford 6th Form. Two sessions wereadministered which consisted of a practice session and the main recorded test.(Materials)For the experiment, the congruencyof the stimuli shown to the participants was the independent variable and thedependent variable was the accuracy and reaction time for the responses. Theaccuracy and time was recorded by a device automatically using software onto aspreadsheet for greater accuracy.The device used to conduct theexperiment was a Lenovo ThinkPad laptop which was running PyschoPy (Version 2x).The software used to conduct the experiment was a python program which waslinked to an Excel spreadsheet that recorded results. The spreadsheet was alsoused to work out the averages of the large set of results.In total, four different stimuliwere used in order to test the processing of local details.
The characters Hand S were used in order to test this in congruent and incongruent forms.However, the format of the stimuli changed randomly and included a large S andH consisting of smaller formations of the same letters. Congruent formsconsisted of the larger letter consisting of the same smaller letter, meaningthat the global and local stimuli were the same. On the other hand, theincongruent form comprised on different global and local stimuli, includingboth S consisting of H and H consisting of S.
(Design)Before the test, the subjects were given the opportunity topractise through a series of preparation questions that allowed them tofamiliarise themselves with the experiment. After this the test followed in theformat mentioned below.A fixation cross appeared in the centre of the screen for 1second, each time, in order to ensure that the subjects visual field was at thecentre of the screen and that they were prepared. Then, the stimulus appearedfor 200 milliseconds at a random point on the screen before beingbackward-masked for 5 seconds. Backwards masking was used in order to ensurethat no trace of the image was left in the subject’s mind as well as on the screen.The stimulus appeared in any of the 4 quadrants at random, with no hint orindication as to where the next stimulus would appear. The subjects had 7seconds to respond to the appearance and to press the corresponding button ofthe smaller letters which formed the larger letter.
During the test, the subjects were exposed to the stimuli 64times in total. The 2 different types of congruent and incongruent forms wereshown 16 times each. The position of the stimuli varied every time in order toprevent the subject from devising a strategy and predict the positioning of thestimuli as this would invalidate the research.The subjects responded using the S and H keys on thekeyboard and were told that their reaction time would be recorded; but thecandidates were given a 7 second time limit to respond. This was done using anExcel spreadsheet which also allowed us to work averages and produce visualinterpretations of the data in the form of a graph. (Procedure)Theexperiment was taken in a regular school classroom, in the presence of 5 othermembers for the first experiment and 4 other members for the second experiment.The participant conducted the experiment on a table, with the laptop positionedcentrally, just below eye level but in a comfortable position for the user.
The shortpractice session at the start of the experiment was conducted in a relaxedenvironment where the participant was able to talk to others and ask for adviceon how to perform the test. The second part to the experiment was done with nointerference with the subject directly. However, the other members in the roomwere free to have conversations amongst themselves. This, as well as otherfactors would have had a effect on the outcome of the experiment and shouldhave been controlled for more accurate results.
(Data)The results gathered from the experiment were recorded ontoa dedicated spreadsheet. The computer program on the laptop measured the typeof stimulus that was shown on the screen, the accuracy and the reaction time. Thestandard deviation of the averages was calculated to allow us to compare thestimuli and accuracy.
The outliers in the experiment were included in the datacollection and the averages, despite the fact that they may have had a smallinfluence on the averages and the standard deviation.(RESULTS)From the data gathered in the experiment, the mean andstandard deviation was worked out. Also, P-values were calculated which is themarginal significance in a hypothesis, showing the degree to which the results weresignificant.The results for the averaged accuracy are shown below in Figure 1; represented in the form of a bar chart and the reaction time isalso included in a table in Table 1.The mean accuracy is greater for congruent stimuli when compared to incongruentstimuli. For congruent stimuli, the percentage accuracy was 84.7% (SD +/-16.
3%)whereas, on the other hand, for incongruent stimuli, the percentage accuracywas only 79.9% (SD +/- 18.5%). This meant there was a 4.8% greater accuracy forthe congruent similarity.
Table 1 – Percentage accuracy of the responses as well as the average reaction time (RT – milliseconds) for both the congruent and incongruent stimuli in the form of a table. Figure 1 – Percentage accuracy of the responses for both the conflicting and consistent stimuli in the form of a bar chart When presented in the form of a bar chart in Figure 1, the variance seems to bequite significant. However, the results achieved from a T-test, to calculatethe P-value, shows that the accuracy difference between the consistent and inconsistentstimuli recognition is not substantial. The P-value (p = 0.528), as shown in Table 2 indicated to us that p>0.05,revealing that the difference is not momentous.
On the other hand, the P-value of the reaction time (p =0.015), shown in Table 2, betweenthe congruent and incongruent stimuli is less than the standard value of 0.05.We can see from this that there is a significant difference for the reactiontime for the two stimuli. The average values obtained for the reaction time hada difference of 62.3ms. As shown in Table1, the incongruent stimuli had an average time of 635.
5ms (SD +/- 119.1ms),whereas for the congruent stimuli, it was faster at 573.2ms (SD +/- 136.2ms).We also calculated the difference in accuracy and reactiontime between the left and right of the screen. This gave us an insight into theidea of left or right dominance of the hemispheres. The variance was notsignificant for both variables as p>0.
05 and the difference between theaccuracy was only 2%; greater on the right-hand side (83.3%) of the screen whencompared to the left-hand side (81.3%). As for the reaction time, 613.6ms forthe left-hand side compared to the 595.1ms for the right. This was only amarginal difference which would be expected on a small-scale experiment. Table 2 – P-values indicating the significance of the different variables.
*Highlighted value for the reaction time shows that the congruent and incongruent stimuli have a significant effect on this dependent variable as it has a value below 0.05. Figure 2 – Reaction time (milliseconds) for both the conflicting and consistent stimuli represented as a bar chart. (Discussion)Before the experiment, we collectively predicted that thereaction time and the accuracy would be greater and faster for congruentstimuli. This was due to the fact that it would be easier for the subjects toprocess the global image first if it consisted of the same information on alocal level, and on the whole, it would decrease the time needed to process theinformation. Consequently, the results support our hypothesis and alsoindicate to us that the results that Navon acquired were reproducible. Asmentioned in his paper, Navon’s data indicated that “global difference are morefrequently detected than local differences”, (Navon, 1977). The data collected by us alsosupports this statement, nonetheless to a smaller extent.
The reaction timesdiffered for the congruent and incongruent stimuli; the congruentrepresentations being faster of the two. The accuracy indicated the sameprinciple.Nevertheless, the experiment had many limitations that had agreat influence on the results achieved.
First of all, the procedure ofrecognising letters consisting of smaller letters on a device not a commoneveryday task. This means that it is not possible to use this research to forma representation of the process of visualisation. To improve the findings, we would have toresearch common tasks that specify the visual processing of the brain. The experiment was also conducted on a very small scale.Using only 11 students may give us a small trend to form the basis of ourconclusion on, but it is not representative of a larger population.
To improvethe accuracy of the experiment, it would have to be conducted on a much largerscale. This would also allow us to find any anomalous values within theresults; which were omitted in this research. References:· Chamberlain,McManus, Riley, Rankin, Brunswick. (2013) –Local processing enhancements associated with superior observational drawingare due to enhanced perceptual functioning, not weak central coherence, 66, (7), pg. 1448-1466.· Gibson.
(1972) – A theory ofdirect visual perception, in J. Royce, W. Rozenboom (eds). The psychology ofknowing. New York: Gordon and Breach.
· Gregory. (1970) – The intelligenteye. London: Duckworth.· Kandel, Eric R, Schwartz, James H, Jessell, Thomas M.(2000) – Principles of Neural Science, (4), pg. 324.· Lamb, Robertson, Knight.
(1989) – Attention andinterference in the processing of global and local information: effect ofunilateral temporal-parietal junction lesions, 27, (4), pg. 471-483.· Navon (1977) – Forest Before Trees: The Precedence ofGlobal Features in Visual Perception. CognitivePsychology, 9, (3), pg. 353-383.
· McLeod (2008) – VisualPerception Theory – https://www.simplypsychology.org/perception-theories.html.
· 1 https://www.brainhq.com/brain-resources/brain-facts-myths/how-vision-works.
· 2 The Asymmetric BrainScholar’s Program Booklet – pg. 34.