Abstract

The Halstead Category Test-Computer Version (HCT-CV) was developed with the goal of adapting the HCT to the computer while maintaining all essential features of the original test. We compared the HCT-CV results from clinical neuropsychological evaluations of 25 patients who were matched on a pairwise basis with 25 patients previously tested with the original version. Matching was done on age, education, and diagnosis. Results of group comparisons showed that the HCT-CV performed comparably with the original version, with mean score differences of <2 points. Correlations with the other subtests from the Halstead–Reitan Neuropsychological Test Battery were also found to be comparable. Results suggest that the HCT-CV can be considered a satisfactory and comparable version of the HCT, while providing advantages in terms of ease of use and portability.

Introduction

The Halstead Category Test (HCT) is one of the most sensitive measures of brain function (Reitan & Wolfson, 1993) and is widely used in neuropsychological assessment. The test was first developed in Dr. Ward Halstead's lab, apparently as an outgrowth of the work of K. Goldstein and M. Scheerer on abstract thinking (Goldstein, 2009) and originally involved sorting or grouping of objects. Halstead and Settlage (1943) reported on an “apparatus” that allowed the presentation of multiple choice items to subjects; they found that brain-damaged subjects performed differently than normal controls in terms of their “grouping behavior.” Halstead (1947) described the “Halstead Category Test (2010)” that presented a series of pictures to a subject seated before a display apparatus. A photograph shows a large box with a small screen displaying a picture of geometric designs. Except for the examiner's panel, which is a small control box in the photo, the apparatus otherwise looks similar to the familiar “gray box” version of the HCT. Halstead's idea about the Category Test was that it involved abstraction and the concept of “oddity” in each of the nine subtests. Subsequent work, especially by Reitan, shortened the test to seven subtests, and this version is the one that has been in wide use for over 50 years. Choca, Laatsch, Wetzel, and Agresti (1997) succinctly summarized the first 50 years of work with the HCT. In this paper, we will refer to the Reitan 7-subtest version, using a gray box and projected stimuli, as the “original version.”

There have been a number of revisions, modifications, and alternate versions developed of the HCT (Mitrushina, Boone, Razani, & D'Elia, 2005); however, none has the independent scientific validation of the original version. Most alternate versions were validated by showing a correlational relationship with the original, but few (if any) have independent validation with multiple groups of brain-damaged patients as was done with the original. Patient groups studied with the original Category Test include Alzheimer's disease (Hom, 1992), tumors (Hom & Reitan, 1984), stroke (Hom & Reitan, 1990), normal aging (Reed & Reitan, 1963), traumatic brain injury (Reitan & Wolfson, 1986, 1988), Parkinson's disease (Boll & Reitan, 1970), alcoholism (Fitzhugh, Fitzhugh, & Reitan, 1965), multiple sclerosis (Reitan, Reed, & Dyken, 1971), Huntington's disease (Boll, Heaton, & Reitan, 1974), epilepsy (Reitan, 1974), and learning disabilities (Selz & Reitan, 1979).

The HCT involves problem-solving, reasoning, abstract thinking, attention, and memory. The examinee is shown a series of designs, projected onto a screen. For each series of designs, there is a single, underlying idea which the examinee must discern through hypothesis testing and trial-and-error learning. For each design, the examinee presses a key on an answer panel to indicate their response. Immediate feedback is given as to whether the answer is correct (a bell rings) or incorrect (a buzzer sounds). For the adult version of the test, there are seven subtests; the older children's (ages 9–14) and young children's (ages 5–8) versions have six and five subtests, respectively.

While the HCT has been extensively validated, it has some limitations in terms of ease of use. For example, a large wooden box is used to display the test stimuli, making the apparatus difficult to transport. In addition, it uses slide projectors which are no longer being manufactured. Thus, it was imperative that a new version be developed. It was important that the new version be as similar to the original as possible—for example, lighted stimuli presented on-screen, patient response via a dedicated answer panel, examiner control of stimulus presentation, and bell/buzzer feedback. This paper describes the validation of a computer version (the HCT-CV; Neuropsychology Center, 2010), developed with the goal of incorporating all the salient features of the original version. Test administration and response recording procedures are the same for the HCT-CV as for the original version, the instructions given to the examinee are the same, and the examinee experience is also very similar. To our knowledge, there are no alternative or modified versions of the HCT that match the original to this degree.

Preliminary studies have been completed that compare the HCT-CV with the original version, using college students as examinees. Silk-Eglit, Miele, Gunner, Lynch, and McCaffrey (2012) reported finding the two versions to be equivalent in their study using a sample (N = 29) of undergraduate students. Half of the sample was tested on the HCT-CV version, and half was tested on the original version. The mean number of errors was statistically comparable for all seven subtests as well as for total errors. The differences in mean raw scores between the two versions ranged from 0 to 2 points on the seven subtests, with most differences being <1 point. The difference in mean scores for the total score was 2.97 points.

Woolery, Farcello, Klimas, Thaler, and Allen (2012) also compared the results on the two Category test versions using a sample of undergraduates (N = 46). They found no statistical differences between the versions on any of the subtests or on the total score. They found a mean error difference of 1.05 points for the total error score.

In the present study, we wished to evaluate the success and suitability of the HCT-CV among clinical patients. We compared the HCT-CV versus the original version among patients with similar characteristics (e.g., age, education, diagnosis). We also evaluated the performance of the HCT-CV in terms of its relationship with other tests within the Halstead–Reitan Neuropsychological Test Battery (HRB). It was important, for purposes of clinical use, that the HCT-CV produce similar test inter-relationships. This would serve to retain the neuropsychologist's ability to conduct integrative clinical analysis and interpretation of HCT results with the HRB (Reitan & Wolfson, 1993).

Method

Participants

Once preliminary testing showed that the HCT-CV appeared to provide the same examinee experience and results as the original version, it was put into routine use in our clinical outpatient facility. From the first 35 consecutive adult clinical patients, 25 patients were matched on a pairwise basis with archival data on 25 patients who were tested with the original version. Matching was first done on diagnosis; from available matches, patients were then selected to match further on age and education. The remaining 10 patients could not be matched within acceptable parameters from our database of patients with complete HRB examinations. All patient pairs except five were matched within 3 years of each other on age; all except four were matched within 3 years on education.

The 50 adult patients had a mean age of 64.2 years (SD = 15.4), with an age range from 30 to 91 years. Their mean education was 15.1 years (SD = 2.8) with a range from 10 to 20 years. Mean Full-Scale IQ (FSIQ) was 112.68 (SD = 14.44) with a range from 80 to 140. There were 32 men and 18 women in the sample.

The patients had all been evaluated for possible neurocognitive dysfunction. Most had underlying medical conditions which were likely contributors to their presenting problems. Patients were matched on their underlying medical condition (where applicable) or on the diagnosis assigned once the neuropsychological evaluation was complete. Matching diagnoses were (number of patient pairs): Hypertension (n = 4), Type II diabetes (n = 5), cerebrovascular disease/accident (n = 3), head injury (n = 2), psychiatric disorder (n = 7), normal for age (n = 3), and Alzheimer's disease (n = 1).

Materials

The HCT-CV is presented via a PC computer (either desktop or laptop) using a Windows operating system (XP, Vista, Windows 7, or Windows 8). The original version presents the stimuli via slide projector. Both versions employ a four-button answer panel by which the examinee gives answers to each stimulus slide. In both versions, the examiner advances the slides and records the patient response on the recording form. The stimuli are identical in the two versions, as are the feedback sounds; the bell/buzzer and sound of the slide advance in the HCT-CV are digitized recordings of the sounds from the original HCT.

Procedures

All patients were tested in a clinical outpatient setting by trained technicians; all were given complete HRBs (including the HCT), Wechsler intelligence scales, an academic screening test, and at least one personality questionnaire.

Statistical Analyses

T-test comparisons were made on subtest and total scores from the respective version of the HCT (HCT-CV or Original). In addition, t-test comparisons were made on all HRB subtests between the two groups of patients. Pearson correlations were computed for the two versions of the Category test, as well as for the respective Category Test and each subtest from the Halstead-Reitan Neuropsychological Test Battery; these latter correlation coefficients were statistically compared.

Results

The groups did not differ in terms of age, education, gender, or FSIQ (Table 1). There were no significant group differences on any of the HRB subtests (Table 2).

Table 1.

Demographic comparisons for the two groups of patients

 HCT-CV (n = 25) Original (n = 25) t-test 
Age (mean [SD]) 64.3 (15.8) 64.2 (15.2) 0.93 NS 
Education (mean [SD]) 15.4 (2.6) 14.8 (2.9) 0.21 NS 
Gender (M/F) 17/8 15/10 0.35* NS 
FSIQ (mean [SD]) 110.92 (15.68) 114.44 (13.18) 0.20 NS 
 HCT-CV (n = 25) Original (n = 25) t-test 
Age (mean [SD]) 64.3 (15.8) 64.2 (15.2) 0.93 NS 
Education (mean [SD]) 15.4 (2.6) 14.8 (2.9) 0.21 NS 
Gender (M/F) 17/8 15/10 0.35* NS 
FSIQ (mean [SD]) 110.92 (15.68) 114.44 (13.18) 0.20 NS 

Note: HCT-CV = Halstead Category Test-Computer Version.

*Chi square statistic.

Table 2.

Group comparisons on HRB subtests

Subtest HCT-CV Original t-test 
Tactual Performance Test (TPT) 17.83 (7.08) 22.69 (10.53) 0.03 NS 
 Total Time (min) 
TPT Memory 6.21 (2.50) 5.8 (2.24) 0.66 NS 
TPT Localization 2.5 (2.25) 2.56 (2.04) 0.80 NS 
Rhythm (#correct) 24.84 (4.79) 23.84 (4.60) 0.22 NS 
Speech Sounds Perception 9.52 (7.41) 10.88 (11.07) 0.58 NS 
Test (#errors) 
Tapping-dominant hand 44.2 (9.60) 44.84 (10.71) 0.75 NS 
Tapping-non-dominant hand 39.2 (9.51) 40.48 (8.80) 0.50 NS 
Trails A (s) 51.44 (59.17) 37.8 (19.48) 0.20 NS 
Trails B (s) 121.7 (89.47) 106.0 (68.37) 0.32 NS 
Impairment Index 0.58 (0.34) 0.64 (0.36) 0.19 NS 
Subtest HCT-CV Original t-test 
Tactual Performance Test (TPT) 17.83 (7.08) 22.69 (10.53) 0.03 NS 
 Total Time (min) 
TPT Memory 6.21 (2.50) 5.8 (2.24) 0.66 NS 
TPT Localization 2.5 (2.25) 2.56 (2.04) 0.80 NS 
Rhythm (#correct) 24.84 (4.79) 23.84 (4.60) 0.22 NS 
Speech Sounds Perception 9.52 (7.41) 10.88 (11.07) 0.58 NS 
Test (#errors) 
Tapping-dominant hand 44.2 (9.60) 44.84 (10.71) 0.75 NS 
Tapping-non-dominant hand 39.2 (9.51) 40.48 (8.80) 0.50 NS 
Trails A (s) 51.44 (59.17) 37.8 (19.48) 0.20 NS 
Trails B (s) 121.7 (89.47) 106.0 (68.37) 0.32 NS 
Impairment Index 0.58 (0.34) 0.64 (0.36) 0.19 NS 

Note: HCT-CV = Halstead Category Test-Computer Version. Scores by group (mean/SD).

The scores on the respective Category Test total score and each Category subtest score were found to be comparable (Table 3). As can be seen from the data in Table 3, the differences between groups in mean scores were fewer than 2 points among all the subtests and total scores. The correlation between the two versions of the Category Test (total score) was 0.66 (p < .01).

Table 3.

Group comparison of scores on each Category Test subtest and total score (errors)

 HCT-CV Original t-test 
Subtest I 0.16 (0.62) 0.1 (0.31) 0.58 NS 
Subtest II 1.24 (2.30) 0.6 (0.68) 0.24 NS 
Subtest III 13.6 (11.11) 16.2 (9.36) 0.46 NS 
Subtest IV 13.44 (12.81) 14.65 (11.39) 0.48 NS 
Subtest V 15.16 (6.94) 15.75 (6.32) 0.88 NS 
Subtest VI 11.12 (7.67) 10.4 (7.66) 0.12 NS 
Subtest VII 6.04 (3.53) 6.6 (3.14) 0.73 NS 
Total Category score 60.72 (33.41) 62.36 (28.35) 0.76 NS 
 HCT-CV Original t-test 
Subtest I 0.16 (0.62) 0.1 (0.31) 0.58 NS 
Subtest II 1.24 (2.30) 0.6 (0.68) 0.24 NS 
Subtest III 13.6 (11.11) 16.2 (9.36) 0.46 NS 
Subtest IV 13.44 (12.81) 14.65 (11.39) 0.48 NS 
Subtest V 15.16 (6.94) 15.75 (6.32) 0.88 NS 
Subtest VI 11.12 (7.67) 10.4 (7.66) 0.12 NS 
Subtest VII 6.04 (3.53) 6.6 (3.14) 0.73 NS 
Total Category score 60.72 (33.41) 62.36 (28.35) 0.76 NS 

Note: HCT-CV = Halstead Category Test-Computer Version.

Pearson's correlations between the respective Category score and scores from each subtest from the HRB were found to be statistically comparable between groups (Table 4).

Table 4.

Correlations with the respective Category test score for each HRB subtest

 HCT-CV Original Fisher transformation z-score 
TPT Total time .42 .52 −0.41 NS 
TPT Memory −.79 −.74 −0.42 NS 
TPT Localization −.59 −.54 −0.25 NS 
Rhythm −.77 −.60 −1.04 NS 
Speech Sounds Percep .77 .52 1.43 NS 
Tap – Dominant hand −.66 −.60 −0.34 NS 
Tap – Non-dominant hand −.53 −.51 −0.09 NS 
Trails A .61 .66 −0.27 NS 
Trails B .68 .67 0.06 NS 
Impairment Index .81 .79 0.22 NS 
 HCT-CV Original Fisher transformation z-score 
TPT Total time .42 .52 −0.41 NS 
TPT Memory −.79 −.74 −0.42 NS 
TPT Localization −.59 −.54 −0.25 NS 
Rhythm −.77 −.60 −1.04 NS 
Speech Sounds Percep .77 .52 1.43 NS 
Tap – Dominant hand −.66 −.60 −0.34 NS 
Tap – Non-dominant hand −.53 −.51 −0.09 NS 
Trails A .61 .66 −0.27 NS 
Trails B .68 .67 0.06 NS 
Impairment Index .81 .79 0.22 NS 

Note: HCT-CV = Halstead Category Test-Computer Version.

Discussion

Within this patient population, the HCT-CV appears to be comparable with the original HCT. Not only were the Category total scores and subtest scores comparable between groups, but the HCT-CV was also found to inter-relate with the other HRB measures in a way that is comparable with the relationships found with the original version.

The HCT-CV uses the same testing procedures as the original. This provides a seamless transition to the new version for those examiners who are familiar with the original. The HCT-CV provides some advantages in terms of ease and accuracy of administration. The software does not allow the patient to select a second answer (except on Slide #1, as required), which has always been an issue in test administration, especially with impulsive examinees. Examiner error is also reduced, in that the slides cannot be accidentally advanced too far, and examinee responses are always correctly recorded.

This study, along with two other recent studies on the comparability of the HCT-CV with the original version (Silk-Eglit et al., 2012; Woolery et al., 2012), suggests that the HCT-CV provides a satisfactory and comparable version of the HCT for clinical use.

Conflict of Interest

The authors are employed by The Neuropsychology Center, PC, which manufactures and sells the HCT-CV.

References

Boll
T. J.
Heaton
R. K.
Reitan
R. M.
Neuropsychological and emotional correlates of Huntington's Chorea
Journal of Nervous and Mental Disease
 , 
1974
, vol. 
158
 (pg. 
61
-
69
)
Boll
T. J.
Reitan
R. M.
Psychological test results of subjects with known cerebral lesions and Parkinson's disease as compared with controls
Perceptual and Motor Skills
 , 
1970
, vol. 
31
 pg. 
824
 
Choca
J. P.
Laatsch
L.
Wetzel
L.
Agresti
A.
The Halstead Category Test: A fifty year perspective
Neuropsychology Review
 , 
1997
, vol. 
7
 (pg. 
61
-
75
)
Fitzhugh
L. C.
Fitzhugh
K. B.
Reitan
R. M.
Adaptive abilities and intellectual functioning of hospitalized alcoholics: Further considerations
Quarterly Journal of Studies of Alcoholism
 , 
1965
, vol. 
26
 (pg. 
402
-
411
)
Goldstein
G.
Neuropsychology in New York City (1930–1960)
Archives of Clinical Neuropsychology
 , 
2009
, vol. 
24
 (pg. 
137
-
143
)
Halstead Category Test – Computer Version [software]
2010
Dallas, TX
Neuropsychology Center
Halstead
W. C.
Brain and intelligence: A quantitative study of the frontal lobes
 , 
1947
Chicago
University of Chicago Press
Halstead
W. C.
Settlage
P. H.
Grouping behavior of normal persons and persons with lesions of the brain
Archives of Neurology and Psychiatry
 , 
1943
, vol. 
49
 (pg. 
489
-
506
)
Mitrushina
M.
Boone
K. B.
Razani
J.
De'Elia
L. F.
Handbook of normative data for neuropsychological assessment
 , 
2005
2nd ed.
New York
Oxford University Press
Hom
J.
General and specific cognitive dysfunctions in patients with Azheimer's disease
Archives of Clinical Neuropsychology
 , 
1992
, vol. 
7
 (pg. 
121
-
133
)
Hom
J.
Reitan
R. M.
Neuropsychological correlates of rapidly vs. slowly growing intrinsic cerebral neoplasms
Journal of Clinical Neuropsychology
 , 
1984
, vol. 
6
 (pg. 
309
-
324
)
Hom
J.
Reitan
R. M.
Generalized cognitive function in stroke
Journal of Clinical and Experimental Neuropsychology
 , 
1990
, vol. 
12
 (pg. 
644
-
654
)
Reed
H. B. C.
Reitan
R. M.
A comparison of the effects of the normal aging process with the effects of organic brain damage on adaptive abilities
Journal of Gerontology
 , 
1963
, vol. 
18
 (pg. 
177
-
179
)
Reitan
R. M.
Vinken
P. J.
Bruyn
G. W.
Psychological testing of epileptic patients
Handbook of clinical neurology, the epilepsies
 , 
1974
New York
North Holland Publishing
Reitan
R. M.
Reed
J. C.
Dyken
M. L.
Cognitive, psychomotor, and motor correlates of multiple sclerosis
Journal of Nervous and Mental Disease
 , 
1971
, vol. 
153
 (pg. 
218
-
224
)
Reitan
R. M.
Wolfson
D.
Traumatic brain injury: Pathophysiology and neuropsychological evaluation
 , 
1986
Tucson, AZ
Neuropsychology Press
Reitan
R. M.
Wolfson
D.
Traumatic brain injury: Recovery and rehabilitation
 , 
1988
Tucson, AZ
Neuropsychology Press
Reitan
R. M.
Wolfson
D.
The Halstead-Reitan Neuropsychological Test Battery: Theory and clinical interpretation
 , 
1993
2nd ed.
Tucson, AZ
Neuropsychology Press
Selz
M.
Reitan
R. M.
Comparative test performance of normal, learning disabled, and brain-damaged older children
Journal of Nervous and Mental Disease
 , 
1979
, vol. 
167
 (pg. 
298
-
302
)
Silk-Eglit
G. M.
Miele
A. S.
Gunner
J. H.
Lynch
J. K.
McCaffrey
R. J.
A comparison of the standard Category Test with a new computer version
2012
Presented at the annual meeting of the American College of Professional Neuropsychology
March 2012
Las Vegas, NV
Woolery
H.
Farcello
C.
Klimas
N.
Thaler
N. S.
Allen
D. N.
Comparability of the computerized and original versions of the Halstead Category Test
2012
Presented at the 32nd annual conference of the National Academy of Neuropsychology
7–10 November
Nashville, TN

Author notes

Partial results from this research were presented at the annual meeting of National Academy of Neuropsychology, 9 November 2012, Nashville, TN, USA.