Even a quick look at this book would indicate that the largest areas of development have been in the construction of new tests and procedures. That is not true for the various tests associated with Arthur Benton. Additional and updated normative information has been made available (Sivan, 1992), but there have been no new tests or procedures. All the same, the Benton procedures continue to be used both clinically and in research settings. The continued utilization of the tests reflects the sound design and the fact that not many tests have been developed focusing on the functions of the right hemisphere.
Arthur Benton has made several large contributions to the field of neuropsychological assessment in terms of test constructions, the understanding of right-hemisphere function, in training, and in professional development. Benton's approach to test construction is to take laboratory tasks that had been found to be sensitive to differences in cortical integrity and to apply them in a clinical setting. Although not all of Benton's tasks actually derive from the laboratory, the construction of all ofthe tasks has the spirit oflaboratory investiga- tions. In addition, the procedures are well defined and standardized. Scoring is described explicitly, and the results are summarized in a quantitative score. A third characteristic of Benton's tests is the use of norms that represent the influence of age and education. Benton was perhaps the first clinical neuropsychologist to have paid more than mere lip service to the confounding effects of age and education in the assessment situation. Finally, Benton was not bound by the left-hemisphere bias of traditional neuropsychology. Many of his tests are either sensitive to the effects of right-hemisphere impairment or, as in the case of the Tactile Form Perception Test, are designed to partial out the linguistic (left-hemisphere) effects from the more purely perceptual (right-hemisphere) effects.
Benton may be the person in this country most responsible for the rise of the flexible approach to neuropsychological assessment. His tests do not constitute a battery. Instead, the tests to be used in each assessment are chosen on the basis of presenting complaints, the type of referral question, or the results of previous testing. Therefore, a different set of assessment techniques is likely to be used on each subject. The tests were not normed on the same set of normative subjects, and therefore comparisons of scores across tests of the same individual are limited. However, the tests are extremely useful for the assessment of particular functions or when administered in conjunction with other assessment devices.
The tests included in the Benton conglomeration are the Benton Visual Retention Test;
the Multilingual Aphasia Examination; and the tests of Temporal Orientation, Right-Left Orientation, and Serial Digit Learning. Benton's perceptual and motor tests include Facial Recognition, Judgment of Line Orientation, Visual Form Discrimination, Pantomime
153
Recognition, Tactile Form Perception, Finger Localization, Phoneme Discrimination, Three Dimensional Block Construction, and Motor Impersistence. These tests are consid- ered singly.
RIGHT -LEFT ORIENTATION
Right-left orientation has been interpreted as indicative of spatial impairment and as being associated with linguistic impairment. The Benton Test of Right-Left Orientation requires the patient to respond with rudimentary motor responses, a requirement that minimizes the influence of aphasic symptoms on the test results. The test consists of 20 commands that are presented in increasing order of complexity. The first set of commands requires the patient to point to right or left parts of his or her own body. The second set of commands requires the patient to point to lateralized parts of his or her body with either the right or the left hand. The third set of commands requires the patient to point to lateralized parts of the examiner's body. Finally, the fourth set of commands requires the patient to point to lateralized parts of the examiner's body with the right or left hand.
There are alternate forms of the test. Form B is identical to Form A except that the words right and left are interchanged. The responses are scored for whether the patient uses the correctly lateralized body part or the correct body part. In addition, errors are separated on the basis of whether the command required the subject to point to his or her own body or to the body of the examiner. Formal investigations of the alternate-forms reliability of the test have not been conducted. However, because of the similarity between the two forms, equiValency can probably be safely assumed. Form R is for patients who cannot use their right hand, and Form L is for patients who cannot use their left hand. Because Forms Rand L are sufficiently different, the alternate-forms reliability of these two forms needs to be investigated.
The test was normed on a sample of 234 male and female subjects who had no history of brain impairment. Statistical analyses indicated no significant differences due to age, sex, or level of education. Distributions of scores for the impaired subjects are also presented in the manual (Benton, Hamsher, Varney, & Spreen, 1983). The impaired subjects consisted of 34 patients with bilateral brain impairment, 20 patients with right-hemisphere lesions, 20 aphasic patients with left-hemisphere lesions, and 20 nonaphasic patients with left-hemisphere lesions. The interpretation of results is based on the similarity of the type of error made by the patient in reference to the normative data. Although interpretations of Forms Rand L are referred to the normative data, formal evaluations of the validity ofthese two forms need to be conducted.
Knowledge about relationships to other tests (concurrent validity) will help us to understand the nature of the deficit underlying poor performance on this test. Information regarding the internal consistency reliability and the test-retest reliability would also enhance the clinical utility of the test.
SERIAL DIGIT LEARNING
Benton's Serial Digit Learning Test is a standardized form of digit supraspan. (The reader is referred to the discussion of supraspan procedures elsewhere in this book.) One of
the advances in Benton's Test of Serial Digit Learning is that nonnative data are available for different levels of education and for different age groups. Another advance is that scoring is sensitive to the degree of error manifested by the patient. Nearly correct responses (an error involving only 1 digit in the series) are scored 1 point, and perfectly correct responses are scored 2 points.
There are two fonns of this test. Fonn SD9 can be given to patients between the ages of 16 and 64. Fonn SD8 can be given to subjects up to the age of 74. Nonnative data are available for 500 medical patients without a history of brain disease. The quantitative score is converted to different percentiles for the age and the education level of the patient. There were no statistically significant differences due to sex in the nonnative sample. Therefore, there are no corrections for sex in the scoring procedures.
Hamsher, Benton, and Digre (1980) found that, in a sample of 100 patients with diagnosed brain impainnent, use of the Serial Digit Learning Test was more likely to result in a correct diagnosis than was the use of the Digit Span subtest of the WAIS. The increase in diagnostic accuracy was greatest for subjects with bilateral lesions. The increase in diagnostic accuracy was least for patients with left-hemisphere lesions. Despite this differ- ence in accuracy as related to laterality of the lesion, it does not appear that the Test of Serial Digit Learning would be useful in lateralizing brain lesions.
FACIAL RECOGNITION
Benton's Test of Facial Recognition is an evaluation of the patient's ability to recog- nize unfamiliar faces from black-and-white photographs. (The reader is referred to the discussion of other tests of familiar and unfamiliar faces elsewhere in this book.) This test has three parts. The first part requires the subject to identify the face in a stimulus photograph from a choice of six front-view photographs. The second part requires the patient to identify the person in the stimulus photograph from a series of six three-quarter view faces, three of which are of the stimulus face. The third part requires the patient to identify the face in the stimulus photograph from a series of six photographs under different lighting conditions. Again, three of the faces are the same as the stimulus face. The test has a short and a long fonn. The long fonn differs only in having a larger number of items. The photographs are organized into a spiral-bound booklet.
The nonnative data are available in the manual based on a sample of 196 medical inpatients with no history of brain disease, psychiatric disorder, or childhood mental deficiency, and on a sample of 90 nonnal subjects. All 286 subjects were between the ages of 16 and 74. Significant differences due to age were found. Significant differences due to education were found only in the 55- to 74-year-old group. Therefore, a correction factor for age was added for the subjects in the 55- to 64-year-old and the 65- to 74-year-old groups. A correction for education was added for the same two groups.
Nonnative data are also available for patients between the ages of 6 and 14. These data are based on a sample of 266 schoolchildren. The IQ score of all of the children was in the 85 to 116 range as measured by the WISe.
The relationship between the short and long fonns has been investigated by correlating the scores of the items in the short fonn with scores of all of the items. The correlation coefficient for 151 of the control subjects was .88, and the correlation coefficient was .92 for the brain-damaged subjects. Overall, a correlation coefficient of .93 was obtained for the
entire sample of 336 subjects. Based on these data, short-form scores can be converted to long form scores by the use of a table in the manual (Benton et aI., 1983).
After the appropriate age and education corrections have been made, the scores are converted to percentiles. Percentile conversions using age conventions are also available for the child sample.
Hamsher, Levin, and Benton (1979) investigated the performance of brain-impaired subjects on the Test of Facial Recognition. The overall sample of 286 subjects consisted of 196 controls, 23 subjects with right anterior lesions, 36 subjects with right posterior lesions, 15 nonaphasic subjects with left posterior lesions, 5 aphasic subjects without comprehen- sion deficits but with left anterior lesions, 8 aphasic subjects without comprehension deficits but with left posterior deficits, 17 aphasic subjects with comprehension deficits and left anterior lesions, and 27 aphasic subjects with comprehension deficits and left posterior lesions. Poor performance appeared to be related to right-hemisphere lesions as well as to left-hemisphere lesions in the presence of comprehension deficits. Levin, Grossman, and Kelly (1977) found that closed-head injury was associated with poor performance on the Test of Facial Recognition only when patients had been comatose for longer than 24 hours.
There were no relationships with the presence of abnormal EEGs or the presence of depressed skull fracture. Furthermore, the Facial Recognition Test does not appear to be related to subcortical white matter hyperintensity as measured by magnetic resonance imaging (MRI) in healthy older individuals (Tupler, Coffey, Logue, Djang, & Fagan, 1992).
However, the Facial Recognition Test does correlate with P300 latency in auditory evoked potentials (Dywan, Segalowitz, & Unsel, 1992). The relationship of the Facial Recognition Test to physiological variables needs to be better explicated.
Levin and Benton (1977) investigated the performance of 44 psychiatric patients whose symptoms implied the presence of cortical impairment and found that their scores were all in the normal range, with the exception of the score of one paranoid schizophrenic.
The authors therefore recommended the Test of Facial Recognition as being useful in differential diagnosis. However, these recommendations need to be tempered in light of the results of Kronfol, Hamsher, Digre, and Waziri (1978), who found that, of 18 severely depressed psychiatric inpatients, 3 performed in the impaired range.
All aspects of the reliability of this test need to be evaluated. In addition, it would be instructive to compare the internal reliability of the short and long forms. More information is needed regarding the validity of this test in psychiatric populations. The relationship of receptive aphasic symptoms to performance on the Test of Facial Recognition needs to be more clearly delineated.
JUDGMENT OF LINE ORIENTATION
The Test of Judgment of Line Orientation (JLO) resulted from Benton's earlier investigations into the effects of right-hemisphere lesions on spatial orientation skills. His earlier investigations involved the tachistoscopic presentation of two lines that mayor may not differ in angle. Because of the difficulty in using tachistoscopic presentation methods at the bedside, a booklet form of this procedure was devised in which the stimulus items are presented as incomplete lines to approximate the level of difficulty when full length lines are presented tachistoscopically (Benton, Varney, & Hamsher, 1978).
There are two fonns of the test, which actually contain the same items presented in slightly different order. There are 30 items in this test. Each item contains two lines of angles varying from the horizontal. For each item, there is also the same template stimulus of a collection of lines in an arrangement with 18-degree increments. The task presented to the patient is to identify the lines in the constant template stimulus that match the angles of the two incomplete lines.
The nonnative data are available in the manual based on a sample of 137 adult control subjects who ranged in age from 16 to 74 years. Significant differences due to age and sex were found. Therefore, corrections are suggested on the basis of these two variables. There was no systematic effect of level of education, so no corrections for this variable are suggested. However, a larger sample may result in systematic differences due to level of education, and this possibility needs to be investigated. Nonnative data are also available based on a sample of 221 children aged 7 to 14 years. Corrected scores are converted to percentiles for the purpose of interpretation. Additional nonnative infonnation is available stratified by age and corrected for level of education is a sample of 750 nonnal adults between the ages of 55 and 97 years (Ivnik, Malec, Smith, Tangalos, & Petersen, 1996).
The split-half reliability of Fonn H, corrected for attenuation, was found to be .94 in a sample of 40 subjects. The split-half reliability ofFonn V was found to be .89 in a sample of 124 subjects. The split-half reliability of the test (the fonn used was not identified) was found to be .84 in a sample of 221 children (Lindgren & Benton, 1980).
Although test-retest reliability has not been evaluated separately for the two fonns, it was evaluated when first one fonn and then the other was administered to a sample of 37 patients. The interval varied from 6 hours to 21 days. The reliability coefficient was .90, and there were no significant differences between the two scores.
Short fonns of the JLO have been developed. Vanderploeg, LaLone, Greblo, and Schinka (1997) reported the results of investigating the psychometric properties of four short IS-item fonns of the JLO. The odd-item and even-item versions possessed adequate reliability. Doubling the scores obtained from the two short fonns resulted in mean scores and standard deviations similar to those in the original fonn in a sample of 81 patients.
Adequate psychometric properties are necessary but not sufficient to recommend clinical use of the short fonns. Woodard et al. (1996) reported the correlations between both the original and the short fonns of the JLO and WAIS-R IQ values, Block Design scores, and Mattis Dementia Rating Scale scores. There were no significant differences among the correlations in comparisons of the short fonns with other and with the original fonn. In addition, classifying a sample of 386 patients using the JLO categories of impainnent from the manual resulted in good agreement between each of the short fonns and the original fonn. Weingarden, Yates, Moses, Benton, and Faustman (1998) examined different short fonns of the JLO, and based on examination of internal consistency as well as correlation with the full fonn, recommended a short fonn composed of items 11 to 30 of Fonn V for clinical use. Woodard et al. (1998) examined the distributions of scores on odd- and even- item versions of Fonn V of the JLO in a sample of healthy elderly subjects. They found nearly equal mean scores and equivalent distributions as well as equivalent relationships with the Beery Visual Motor Integration test and suggested that these two short fonns be used in conducting screening evaluations or serial evaluations of the elderly.
The sensitivity of the test to lateralized impainnent was investigated in a sample of 50 subjects with left-hemisphere lesions and 50 subjects with right-hemisphere lesions.
None of the subjects were aphasic. Although both groups contained subjects with a wide range of scores, the right-hemisphere 1esioned subjects were much more likely to perform in the impaired range. The test therefore appears to be sensitive to 1ateralized impairment.
However, much more knowledge is needed regarding the construct validity of the test.
Preliminary investigations of this type are promising. Benton et al. (1983) reported that, at least for children, there is a low partial correlation of this test with the Test of Facial Recognition (when age and Verbal IQ are corrected for), indicating that the two tests mea- sure separate constructs. As another way of examining the construct validity of the JLO, Kempen, Kritchevsky, and Feldman (1994) investigated the impact of visual impairment (refractive error) on performance on the JLO and found that although visual impairment affected scores on the Facial Recognition and Visual Form Discrimination tests, there was no effect on the JLO.
Trahan (1998) reported that twice as many right cerebrovascular accident (CVA) patients exhibit impaired performance on the JLO as left-hemisphere CVA patients. This result was supported by the findings of York and Cermak (1995) in a smaller sample of patients with either right- or left-hemisphere CVA. Another method of investigating the laterality of JLO performance involves the use of functional imaging techniques in non- clinical subjects. Deutsch, Bourbon, Papanicolaou, and Eisenberg (1988) reported that regional cerebral blood flow shows right-sided asymmetry when 19 subjects were adminis- tered the JLO. Hannay et ai. (1987) reported similar results when specifically looking at the right temporo-occipital region. Therefore the JLO appears to be related to right-hemisphere activity in the noninjured brain. Scopolamine, a drug that interferes with cholinergic activity, has a negative effect on JLO performance (Meador et aI., 1993).
Although the JLO has norms available for both adults and children and has been suggested for use in both sets of subjects, there may be differential validities in the two age groups. Riccio and Hynd (1993) reported that although the JLO correlates with the visual- spatial measures of the WISC-R and with math ability (WRAT-R), the JLO did not differentiate among groups of psychiatric, learning disabled, and normal control children.
Yet another way to interpret results of the JLO is related to types of errors exhibited by the subjects. Ska, Poissant, and Joanette (1990) reported that certain types of errors were specific to patients with Alzheimer's dementia. However, another set of researchers failed to replicate this finding and instead reported that patients with Parkinson's disease exhibited a specific set of errors types (Finton, Lucas, Graff-Radford, & Vitti, 1998). This second set of results is in need of replication, but the approach of error analysis holds promise for future applications of the JLO.
VISUAL FORM DISCRIMINATION
The Test of Visual Form Discrimination (VFD) is an attempt to assess the ability of a patient to perceive and recognize complex visual patterns. The test stimuli are spiral-bound in a booklet. Each item consists of a target stimulus and four choices, one of which matches the target exactly. Each stimulus consists of two major figures and a peripheral figure. Aside from the correctly matching stimulus, one of the alternate stimuli involves the rotation of a major figure, one involves the rotation of the peripheral figure, and one involves the distortion of a major figure. Qualitative information about the type of error made can be derived from this multiple-choice test.