Free Health Essays - Critical evaluation of the research article "Holistic and part face recognition in children with autism" demonstrating understanding of aims, sampling, data collection methods, analysis strategies, validity and potential application and ethical considerations.
In order to try and improve quality of life and social interaction for those with the condition, social and biological aspects of communication have been researched in order to attempt to uncover what is inhibiting communication ability, and whether this can be reversed or adapted, in order to regain a greater degree of normal functionality. One particular aspect that has become popular within the field is investigation into face-recognition or more specifically, whether the ability to recognize faces is impaired in individuals with the above conditions.
Human nature inclines us towards faces – newborns have been found to distinguish between their mother and a stranger, so it appears we hold an innate readiness to recognize the faces of those we know or have met before. Human ability to discriminate between massive numbers of faces to recognize people we have met or seen before is quite astounding – despite the large number of people we see in a lifetime, and considering the relative similarity of the human face (most of us have two eyes, a nose and a mouth in standard placing on the face) even the smallest differences can be picked up on.
Recognition of people (face-recognition) is thus important for social development in terms of personal identification ability and reciprocal communication episodes so in cases where this is impaired, social development can also be affected. Autistic individuals have been shown to pay attention to different aspects of the face compared to non-autistic individuals (Pelphrey et al, 2002), both in communication instances (lack of eye contact) and according to the literature, in the facial-recognition process, with the finding that autistic individuals pay particular attention to the mouth region. Problems in facial-recognition ability becomes apparent from infancy (Corsello, 2005) with autistic children found to be delayed in demonstrating early ‘face-related milestones’ such as looking into another persons face or observing the reactions of others, and therefore such problems in infancy can act as an initial screening tool for autism (Ballie, 2001).
This study was performed from the perspective that autistic children are not only delayed in their facial-recognition capacities (as they are in other developmental capacities) but that they perform the task fundamentally different to ‘normal’ non-autistic children. (Joseph & Tanaka, 2003) This assumption has also been demonstrated biologically in terms of areas of the brain activated and in a decrease in processing speed in autistic individuals (McPartland et al, 2004) Scans of brains of autistic individuals have shown reduced activity in areas of the brain that process images of faces, and increased activity of the inferior temporal gyri-an area (responsible for inanimate object processing) during facial-recognition tasks (Schultz et al, 2000), which may indicate that faces are processed in the same way that objects are, and thus abnormal in such use of technique.
Young non-autistic children have been found to process facial information on a part (feature basis) although this method becomes redundant with increasing age when the holistic (entire) method is initiated – This is based on studies which have demonstrated that after such ages, non-autistic children perform significantly worse on inverted facial-recognition tasks (due to the use of the holistic method) than children of a young age (Flin, 1985) or than older adolescents with autism (van der Geest, Kemner, Verbaten & van Engeland) as disruption is less experienced for inverted inanimate objects, which are processed by part-by-part (part-based) recognition in all individuals. Unlike upright face images, inverted images are best identified using the part-based method which is why autistic children will not be different to normal children for these tasks, but will be worse at upright images which are more efficiently processed holistically. Facial recognition is found to be compromised when part-based techniques are used due to the fact that judgment is impaired as the relationships between constituent parts of the face which usually provide individuals with the sensitivity to pick up on the smallest differences between faces, is not there. The use of holistic processing is therefore linked with (normal) development.
These differences in the processing of faces by people with autism forms the basis of investigation for the current study which set out to investigate the specifics of this difference by testing the premise that it is an impairment in holistic face-recognition processes (where the face is processed in its entirety) by relying to an unusual extent on part-based face recognition (feature by feature separately) strategies that leads to a decrement in ability for autistic individuals (Joseph & Tanaka, 2003) which to date (of the article) had yet to be tested directly.
The procedure set out within the method that was used for both studies within the paper was taken from an earlier study by Tanaka et al (1998) entitled the whole-part method, which through a series of photographs, the participants were asked to identify from two pictures on a computer screen, which they had seen before. Depending on whether the task block was testing whole or part-based technique, a photograph of a whole face with a doctored feature in the foil picture out of the two on screen (whole face testing) or pictures of two different pictures of a single feature (i.e. two different people’s noses) were presented. The children’s answers were recorded by the computer via their response to the touch-screen for each question as to which face they had seen before. This particular method benefits from the fact that it has been used successfully on typically developing children in an earlier study, and therefore was not required to be modified to the age bracket the study was involved with (children) as many cognitive tests are originally designed for adults. As the earlier study clearly outlined the constraints with which the application of the methodology was performed from (for example exposure to the first picture of a face the child was then to find was for 5 seconds etc) the authors therefore had the benefit of using a test that had already been checked in its reliability and validity of measures. However such benefits to its use are subject to three main considerations which should be noted. First of all, the current study contains a subject group of not only typically developing children (as in Tanaka et al 1998) but an autistic group and comparison group of children with a history of language difficulties and delay. Therefore the reliability of a test with a somewhat complicated procedure has been checked for its standard in relation to typically developing children, not children with language and/or developmental problems such as for the subjects within the autistic and comparison group – such an issue can be seen as evident due to the drop out rates of subjects through not being able to complete the training section prior to the experiment: typically developing children had no subjects unable to complete training, the comparison group lost 3 children due to this, and 11 from the autistic group. This factor means that the concurrent validity of the procedural measure for the autistic sample in regards to the earlier typically developing sample, is questionable (Rust & Golobok, 1992). The high numbers of autistic subject drop-outs due to inability to complete training may also incur another issue in regards to the data. It was identified from baseline characteristics of the drop out subjects from the autistic group that these children had more symptoms of autism than those that remained in the study (were able to sufficiently complete training) and therefore it may be the case that only higher functioning autistic children could complete the experiment. From this, that fact that autism as a condition can be seen to also include conditions such as Aspergers syndrome which is very similar to autism but with generally higher functionality (Macintosh & Dissanayake, 2004) is the sample truly autistic, or is it also inclusive of the autistic spectrum disorders and so are differences in processing evident across the board of autistic disorders, or one over another. These questions cannot be answered from the current study, but could be something to consider within the scope of future research. A second deviation from the original methodology by Tanaka et al (1998) is the change from the use of sketches to photographs that were presented for the subjects to identify. Although initially photographs may have the benefit of presenting a more realistic image (better ecological validity), some authors have cited that the cut-and-paste nature of photographs in these type of experiments for whole-face pictures can make the images not ‘pure.’ (Leder & Carbon, 2004) What is meant by this, is that when the photographic image of a feature is taken from one face and superimposed on another the sizing of one feature against the rest of the face may be out (the bridge of a nose from a nose feature may not match up with the eyes of the new face) and due to the fact that holistic processing is thought to involve sensitivity to the smallest of relationships between features, how accurate can these cut-and-paste photo images be for representing the way in which children view living-breathing, human faces.
Lastly the way in which the authors decided to run the experiment meant that some of the children had breaks between trials whilst others did not – children were given the option of a break, and most opted to continue without. This may have meant that influences such as tiredness and loss of concentration due to prolonged time periods without a break may have affected the results for some of the children (task-related error – Dyer 1995) so it is proposed that setting a time break as standard would have been a more reliable procedural requirement than allowing the children the option of taking a break or not.
The subject groups themselves included the results from 50 children involved in Study 1 for typically developing children, with 22 children forming the final numbers after exclusions in the autistic group, and 20 in the comparison group for Study 2. With a total sample of less than 100 participants overall generalisability to the autistic population is compromised, and it is felt that the autistic group should have been significantly larger, with at least equal numbers to the typically developing and comparison group total (Breakwell, Hammond & Fife-Shaw, 1995).
Subjects were allocated to the autistic group if they satisfied the criteria set out by the Autism Diagnostic Interview-R (AID-R) with classification confirmed by a psychologist. The use of a qualified professional and not simply the result of the AID-R may go someway towards minimizing misclassification of children with disorders similar to autism (e.g. Aspergers) as it is not discussed in the article whether the AID-R is capable of distinguishing between the different disorders on the spectrum of autism, although the professional psychologist may have been able to pick up on this.
The precise method of sampling in which the autistic group was recruited is not given in the article, other than the vague statement that they were recruited from community sources. This therefore has the potential to bias the sample depending on whether autistic children were sampled from a range of areas geographically and in relation to the autistic community, and also depending on what method of sampling was employed.
The baseline characteristics about age IQ and AID-R scores for the autistic and comparison groups were taken, with age and IQ found not to be significantly different from each other, but with the comparison group scoring significantly less on the AID-R score (and hence group allocation) Demographic information about the sample was limited to stating the location for the typically developing group, and this is felt to be a limiting factor which could have impacted the results via participant-related error (Dyer,1995) due to the fact that variables such as SES (socio-economic status, parental education level etc) can have a strong effect on the results found.
Analyis involved the use of different statistical analyses (ANOVA’s and T-tests) and the results of these found that typically developing children (Study 1) were better at recognizing face parts represented in the whole face than in parts in all the upright, but not inverted faces. Features included in the analysis of part-based procedure were changed after the initial analysis results were calculated. Originally including all features the results subsequently found an advantage for nose features. As the authors were interested in regards to the differences in attention for features of eye’s and mouths (based on earlier research), the only features included in the second run of the analysis were the data from eye’s and mouths, for which eye’s were then found to hold the advantage for normal children (consistent with the hypothesis that normal children had a preference to eyes), and this may be seen to have picked variables in order to fall in line with the original hypothesis (that eyes would be the strongest feature for identification on non-autistic individuals) in order for the typically developing group to be different to what was expected for the autistic group. The finding of the advantage of a new feature may have been an important finding that the authors have subsequently ignored in order to match up to the a priori hypothesis, based on existing research – it might have been more interesting to have found something new, than to merely add to a finding already established (priority of eye’s over mouths).
An interesting finding for the autistic group is that they did in fact indicate an advantage in upright whole face recognition and not in inverted faces which is consistent with a holistic processing technique, but only for tasks that included the mouth feature as target. The analysis between the studies thus found an advantage in autistic over typical children in recognizing differences between faces from pictures of the mouth, but that the autistic group was far less efficient than the typically developed group of children at distinguishing between faces by the eyes. Analysis from the comparison group of non-autistic but language-delayed children, found the results to be closer to the typically developing children than to the autistic children. The use of the comparison group to test whether they were closer to typically developed or autistic children proved an interesting inclusion, in highlighting that there is something specific to autism and not language delay, for the impairments in face-recognition.
The findings of the study are acknowledged as adding to the literature on the topic of face-recognition capacity for people and children with autism, with the indication that a complete lack of ability appears not to be the case. However the generalisability of the study is somewhat limited by the small sample numbers, and a range of questions regarding the suitability of the procedure to lower-functioning autistic children (predominantly higher-functioning autistics comprised the final autistic sample), and sampling reliability through a lack of background information provided about sampling methods and the comparability of general demographic data between all of the groups, so to ensure sample diversity in regards to race, SES and family backgrounds - the current study was conducted within a single state so such factors are likely to have compromised.
The authors thus conclude that the results of the study back up much of the literature in regards to impairment rather than complete inability of autistic children to process faces holistically. The importance of the mouth in recognition tasks is reflected as a possible consequence of autistic individual’s general aversion to eye contact with others, or alternatively could be seen as a reason why autistic people avoid eye contact, as they may feel more comfortable concentrating on the mouth region of the face or use observation of the mouth for lip-reading. Such information could therefore be used for guidance on methods of communication with autistic individuals, and in helping to educate parents on why this behavior (lack of eye contact) may be occurring, which can be upsetting to many. The authors should consider in their plans for future research developing a methodology which is more suited to the communication restraints involved in working with autistic subjects (to lessen the drop-out rate of autistic subjects due to not being able to complete training)
References
Ballie, R (2001) ‘Face recognition ability could be early screen for autism’ Monitor on Psychology Vol 32, No 7
Banyard, P & Grayson, A (2000) Introducing Psychological Research (2nd Ed) Palgrave
Bartlett & Searcy (1993) cited in Joseph, R. & Tanaka, J. (2003) ‘Holistic and part-based face recognition in children with autism’ Journal of Child Psychology and Psychiatry Vol. 44 Issue 4
Breakwell, G. Hammond, S. & Fife-Shaw, C (1995) Research methods in psychology Sage Publications
Carey & Diamond, (1977), in cited in Joseph, R. & Tanaka, J. (2003) ‘Holistic and part-based face recognition in children with autism’ Journal of Child Psychology and Psychiatry Vol. 44 Issue 4
Coolican, H (1999) Aspects of Psychology: Research methods and statistics Hodder & Stoughton
Coolican, H (1999) Research methods and statistics in psychology (3rd ed) Hodder & Stoughton
Corsello, C (2005) ‘Early intervention in autism’ Infants & Young children Vol 18. No. 2
Dyer,C (1995) Beginning research in Psychology: A practical guide to research methods and statistics Blackwell
Flin (1985) cited in Joseph, R. & Tanaka, J. (2003) ‘Holistic and part-based face recognition in children with autism’ Journal of Child Psychology and Psychiatry Vol. 44 Issue 4
Freire & Lee, (2001) cited in Joseph, R. & Tanaka, J. (2003) ‘Holistic and part-based face recognition in children with autism’ Journal of Child Psychology and Psychiatry Vol. 44 Issue 4
Hobson, Ouston & Lee, (1988) cited in Joseph, R. & Tanaka, J. (2003) ‘Holistic and part-based face recognition in children with autism’ Journal of Child Psychology and Psychiatry Vol. 44 Issue 4
Joseph, R. & Tanaka, J. (2003) ‘Holistic and part-based face recognition in children with autism’ Journal of Child Psychology and Psychiatry Vol. 44 Issue 4
Langdell et al, (1978) cited in Joseph, R. & Tanaka, J. (2003) ‘Holistic and part-based face recognition in children with autism’ Journal of Child Psychology and Psychiatry Vol. 44 Issue 4
Leder, H. & Carbon, CC. (2004) ‘Part-to-whole effects and configural processing in faces’ Psychology Science Vol 46 No. 4
Osterling & Dawson, (1994) cited in Joseph, R. & Tanaka, J. (2003) ‘Holistic and part-based face recognition in children with autism’ Journal of Child Psychology and Psychiatry Vol. 44 Issue 4
Macintosh, K & Dissanayake, C (2004) ‘Annotation: The similarities and differences between Autistic disorder and Aspergers disorder: A review of the empirical evidence’ Journal of Child Psychology & Psychiatry & Allied Disciplines Vol. 45 Issue 3
McPartland, J. Dawson, G. Webb, S. Panagiotides, H & Carver, L (2004) ‘Event-related brain potentials reveal anomalies in temporal processing of faces in autism spectrum disorder’ Journal of child psychology & Psychiatry & Allied Disciplines Vol. 45 Issue 7
Pelphrey, K. Sasson, N. Reznick, S. Paul, G. Goldman, B. & Piven, J (2002) Journal of Autism and Developmental Disorders Vol. 32 Issue 4
Rust, J & Golombok, S (1992) Modern Psychometrics: The science of psychological assessment Routledge
Schultz et al, (2000) cited in Joseph, R. & Tanaka, J. (2003) ‘Holistic and part-based face recognition in children with autism’ Journal of Child Psychology and Psychiatry Vol. 44 Issue 4
Tanaka, J. & Farah, M (1993) ‘Parts and wholes in face recognition’ Quarterly Journal of Experimental Psychology Vol. 46a
Tanaka, J. Kay, J. Grinnell, E. Standsfield, B. Szechter, L (1998) ‘Face recognition in young children: when the whole is greater than the sum of its parts’ Visual Cognition Vol. 5 No. 4
van der Geest, J. Kemner, C. Verbaten, M & van Engeland, H. (2002) ‘Gaze behaviour of children with pervasive developmental disorder toward human faces: a fixation time study’ Journal of Child Psychology & Psychiatry & Allied Discplines Vol 43, Issue 5
