Computerized Neurocognitive Scanning: II
Total Page:16
File Type:pdf, Size:1020Kb
Computerized Neurocognitive Scanning: II. The Profile of Schizophrenia Ruben C. Gur, Ph.D., J. Daniel Ragland, Ph.D., Paul J. Moberg, Ph.D., Warren B. Bilker, Ph.D., Christian Kohler, M.D., Steven J. Siegel, M.D., Ph.D., and Raquel E. Gur, M.D., Ph.D. Cognitive dysfunction in schizophrenia is well established neurocognitive scan and clinical variables, including with neuropsychological batteries, which have assessed premorbid adjustment, onset age, illness duration, quality multiple domains indicating diffuse deficits especially in of life, and severity of negative symptoms. These processing related to frontotemporal systems. Two studies correlations were higher and more prevalent in women than are reported examining the feasibility of the computerized men, who showed correlations predominantly for speed neurocognitive scan to assess differential deficits in rather than accuracy. Neuroleptic exposure was associated schizophrenia. In Study 1, we tested 53 patients and 71 with poorer performance only for speed of memory controls with the traditional and computerized assessments processing, and in men, this association was seen only for counterbalanced in order. Both showed comparable typical neuroleptics. We conclude that the computerized generalized impairment in schizophrenia with differential neurocognitive scan can be applied reliably in people with deficits in executive functions and memory. The profile was schizophrenia, yielding data that support its construct and replicated in Study 2 in a new sample of 68 patients and 37 criterion validity. controls, receiving only the computerized scan. The [Neuropsychopharmacology 25:777–788, 2001] combined sample showed robust correlations between © 2001 American College of Neuropsychopharmacology. performance on both speed and accuracy measures of the Published by Elsevier Science Inc. KEY WORDS: Schizophrenia; Neurocognition; Neuropsycho- 1997; Heaton et al. 1994). Earlier studies, which com- logical testing; Computerized assessment; Sex differences monly focused on one domain such as attention or memory, have not linked cognitive aberrations to brain Generalized cognitive deficits, considered a critical fea- systems (Calev et al. 1983). However, advances in neu- ture of schizophrenia, have been well documented roscience have provided the methodological tools (Blanchard and Neale 1994; Goldstein 1986; Gur et al. needed to examine schizophrenia from a neuropsychi- atric perspective. The availability of standardized neuropsychological batteries, developed and applied in neurological popu- From the Schizophrenia Research Center, Neuropsychiatry Sec- lations, afforded evaluation of the profile of patients tion, Department of Psychiatry (RCG, JDR, PJM, CK, SJS, REG, WBB), and the Department of Biostatistics and Epidemiology (WBB), The with schizophrenia interpretable from the perspective University of Pennsylvania, Philadelphia, Pennsylvania. of regional brain function. Although different tests Address correspondence to: Ruben C. Gur, Ph.D., Brain Behavior were applied, they generally tapped the same neu- Laboratory, Department of Psychiatry, The University of Pennsylva- nia, 10 Gates Bldg., 3400 Spruce St., Philadelphia, PA, 19104. Tel.: ϩ1- rocognitive domains: abstraction and mental flexibility, 215-662-2915; Fax: ϩ1-215-662-7093. E-mail: [email protected] attention, memory, and language. Against a diffuse pat- Received January 11, 2000; revised April 17, 2001; accepted April tern of dysfunction, reflecting reduced brain capacity or 19, 2001. Online publication: 4/26/01 at www.acnp.org/citations/Npp reserve, patients with schizophrenia seemed to show 042601113. relatively greater impairment in executive functions, NEUROPSYCHOPHARMACOLOGY 2001–VOL. 25, NO. 5 © 2001 American College of Neuropsychopharmacology Published by Elsevier Science Inc. 0893-133X/01/$–see front matter 655 Avenue of the Americas, New York, NY 10010 PII S0893-133X(01)00279-2 778 R.C. Gur et al. NEUROPSYCHOPHARMACOLOGY 2001–VOL. 25, NO. 5 and in learning and memory, related to frontotemporal women) from the Schizophrenia Center of The Univer- dysfunction (Gold et al. 1992; Goldberg et al. 1987; sity of Pennsylvania. Healthy participants were se- McKenna et al. 1990; Saykin et al. 1991, 1994). lected from the larger sample presented in the compan- Neurocognitive measures can help address the fol- ion article, to balance patients sociodemographically lowing pivotal questions for understanding the patho- with respect to age and parental education. The groups physiology of schizophrenia. When in the course of the did not differ (mean Ϯ SD) in age (patients 34.1 Ϯ 11.1; illness do cognitive deficits emerge? Are they progres- controls 31.2 Ϯ 9.7) or parental education (patients sive? Do they relate to the symptoms that define the 12.9 Ϯ 3.2; controls 12.2 Ϯ 4.2), but, as expected, patients disorder? Are they affected by treatment and relate to attained lower education (13.4 Ϯ 2.3) than controls outcome? Studies addressing these questions have indi- (15.9 Ϯ 2.0), t(120) ϭ 6.18, p Ͻ .0001. Patients had a DSM- cated that: neurocognitive deficits are apparent at first IV diagnosis of schizophrenia established by medical, clinical presentation in neuroleptic-naive patients neurological, and psychiatric evaluations including (Bilder et al. 1992; Hoff et al. 1991; Saykin et al. 1994); clinical assessment, structured interview (SCID-P, First while stable, some show improvement with treatment et al. 1996), history obtained from family, care provid- (Censits et al. 1997), especially with atypical neurolep- ers, and records. Those with schizophreniform disorder tics (Buchanan et al. 1994; Daniel et al. 1996; Green et al. at study entry met criteria for schizophrenia at follow- 1997; Hagger et al. 1993); and they show limited rela- up. Participants had no history of any other disorder or tion to clinical symptoms (Censits et al. 1997), but do re- event that might affect brain function. Age of onset of late to functional outcome (Bellack et al. 1999; Green psychotic symptoms in the context of functional decline 1996). was 22.5 Ϯ 5.4, and duration of illness was 10.4 Ϯ 9.5 To achieve maximum benefit from neurocognitive years. Patients were clinically stable at the time of study measures in basic and clinical neuroscience, and in in- with mild to moderate symptoms (22 inpatients, 31 out- tervention, it is necessary to use tests that yield efficient patients). There were 28 neuroleptic-naive and 25 previ- and accurate performance data and that can be easily ously treated patients: 13 patients (24% of sample) were applied both in the laboratory setting and in clinical tri- on typical neuroleptics, and 9 (17%) were on atypical als. Traditional paper-and-pencil batteries have several agents at the time of testing. limitations, and we have described in the companion paper a computerized scan that may address them. Study 1: Procedures We present two studies aimed at examining the util- ity of the computerized scan in neurocognitive evalua- Clinical and neurocognitive assessments were con- tion of schizophrenia. In study 1, we applied both the ducted within a week. After complete description of the traditional and the computerized neurocognitive as- study, written informed consent was obtained before sessment to a sample of well-characterized patients participation. Assessments of symptoms and level of with schizophrenia. This enabled a comparison of the function were performed by trained reliable (ICC Ͼ resulting cognitive profiles in relation to sample charac- 0.85) investigators (Gur et al. 1991). Symptom ratings teristics. We hypothesized that the diffuse pattern of included the Scales for Assessment of Negative Symp- deficits evident in patients with the traditional battery toms (SANS, Andreasen 1984a) and Positive Symptoms will also be manifested in the computerized scan. The (SAPS, Andreasen 1984b). Premorbid Adjustment Scale computerized scan permitted, in addition, an evalua- (PAS, Harris 1975), and Quality of Life Scale (QOL, tion of whether patients are differentially affected in Henrichs et al. 1984) assessed functioning. performance accuracy versus performance speed. In The traditional neuropsychological battery described study 2, we extended the computerized scan to a new in the companion paper (Gur et al. 2001) was adminis- sample of patients to establish replicability of the neu- tered to participants according to established proce- rocognitive profile. By combining the two patient sam- dures by trained predoctoral and postdoctoral Fellows ples, we also had the power to examine sex differences (Censits et al. 1997; Ragland et al. 1999,2000; Saykin et and correlations with clinical measures, which help es- al. 1991, 1994). The computerized neurocognitive scan tablish construct and criterion validity. was administered to all participants in a counterbal- anced order within a week of each other. Patients re- mained clinically stable between the administration of STUDY 1. COMPARISON OF TRADITIONAL both tests, and no changes in medications were intro- BATTERY AND COMPUTERIZED SCAN: duced. METHODS Study 1: Subjects Study 1: Data Analysis Participants were 53 patients with schizophrenia (34 Test scores were standardized (z-scores; mean ϭ 0, stan- men, 19 women) and 71 healthy controls (41 men, 30 dard deviation ϭ 1), based on all healthy participants in NEUROPSYCHOPHARMACOLOGY 2001–VOL. 25, NO. 5 Computerized Neurocognitive Scanning: II 779 our normative database, and grouped into summary The hypotheses of impaired functioning