Caudate size in first-episode neuroleptic-naive schizophrenic patients measured using an artificial neural network

Abstract

Background: Structural brain imaging studies have demonstrated an increase in caudate volume in schizophrenic patients medicated with typical neuroleptics and a volume decrease following treatment with atypical neuroleptics. The measurement of striatal volume in patients who have never been treated with neuroleptics may indicate whether these changes are superimposed on intrinsic basal ganglia pathology in schizophrenia or are solely neuroleptic-induced.

Methods: We studied 36 first-episode, neuroleptic-naive schizophrenic patients and 43 control subjects using an artificial neural network (ANN) to identify and measure the caudate nucleus. The resulting volumes were analyzed using an ANCOVA controlling for intracranial volume, age, gender, and socioeconomic status.

Results: The mean volume difference between the caudate nuclei of patients and control subjects was .297 mL, the caudate nuclei of the patients being smaller than those of controls. When we covaried for intracranial volume, this was a statistically significant difference in caudate volume (n = 79; df = 1,75; F = 4.18; p > .04).

Conclusions: Caudate nuclei of neuroleptic naive schizophrenic patients are significantly smaller than those of controls. This suggests that patients suffering from schizophrenia may have intrinsic pathology of the caudate nucleus, in addition to the pathology observed as a consequence of chronic neuroleptic treatment.

Introduction

Like many complex structures in the mammalian forebrain, the basal ganglia have been associated with an increasing number of motor and cognitive tasks during recent years Graybiel 1990, Graybiel 1997, Goldman-Rakic 1990. Reflecting our perplexity concerning their multiple functions, more recent scientists have reminded us of Wilson’s depiction of the basal ganglia as “the dark basements of the brain” Graybiel 1990, Goldman-Rakic 1990, whose structure and function require illumination so as to enhance our understanding of the potentially hazardous results of striatal injury and neuronal misconnection. The basal ganglia have also come to be recognized as a vital component of the corticostriatonigral–thalamocortical circuitry Graybiel 1997, Goldman-Rakic 1990, Selden et al 1994, connecting subcortical to cortical regions. Therefore basal ganglia aberrations are prone to have far reaching consequences. Over the years, a wide spectrum of disorders have come to be associated with loss of basal ganglia integrity. Heimer et al (1991) draws attention to the convergence of disorders such as Parkinson’s disease, Huntington’s disease, and schizophrenia, and conjectures that much of their similarity with regard to symptomatology points to pathologic changes in the striatal domain, since all of these disorders are characterized by the cognitive impairments, abnormal movements, and psychoses indicative of subcortical pathology.

The various components of the striatum (caudate, putamen, globus pallidus, nucleus accumbens, and olfactory tubercle) have highly differentiated neurochemical structure and functions Graybiel 1990, Graybiel 1997, Holt et al 1997, Selden et al 1994. Each striatal component is unique in possessing a characteristic set of afferent and efferent connections within the corticostriatonigral–thalamocortical circuit, segregating functions so that every cortical area has a “private line” through to specific striatal areas and vice versa (Goldman-Rakic and Selemon 1990). The caudate nucleus also has its own specific associative and cognitive functions Graybiel 1997, Goldman-Rakic 1990, Selden et al 1994. Structural and functional insults to the caudate nucleus reportedly lead to motor impersistence, dressing and constructional apraxias, abulia, language deficits, memory problems, and changes in affect Mendez et al 1989, Stein et al 1984. Reciprocal neural circuits connect prefrontal areas to the caudate nuclei; these circuits appear to affect personality, learning, problem solving, and the initiation and inhibition of spontaneous behavior. Malfunction of the caudate nucleus as part of the corticostriatonigral–thalamocortical circuitry is reflected in the observation that psychiatric manifestations of diseases involving the caudate nucleus (such as learning problems found in children with ADHD, inhibition of spontaneous behavior and compulsions in Tourette’s syndrome or obsessive compulsive disorder, and affective and personality changes in Parkinson’s disease and Huntington’s disease) resemble some of the features of patients with purely prefrontal lesions as well as some of those of patients with purely striatal lesions Aylward et al 1994, Aylward et al 1997, Castellanos et al 1994, Doraiswamy et al 1995, Hynd et al 1993, Krishnan et al 1992, Mataro et al 1997, Mattson et al 1996, Peterson et al 1993, Robinson et al 1995, Stein et al 1997, Swayze and Yuh 1992. Volumetric measures of the caudate have been made in many of these disorders, e.g., fetal alcohol syndrome (Mattson et al 1996), attention deficit hyperactivity disorder (ADHD) Castellanos et al 1994, Hynd et al 1993, Mataro et al 1997, Tourette’s syndrome (Peterson et al 1993), obsessive compulsive disorder Robinson et al 1995, Scarone et al 1992, Stein et al 1997, affective disorder Aylward et al 1994, Doraiswamy et al 1995, Krishnan et al 1992, Swayze and Yuh 1992, and Huntington’s disease (Aylward et al 1997). These studies cite caudate volume decrement as attributable to toxin-induced neuronal destruction (Mattson et al 1996), and disease-related degeneration of striatal neurons Aylward et al 1997, Krishnan et al 1992, Peterson et al 1993, Robinson et al 1995. At the opposite end of the spectrum, caudate enlargement has been cited as being due to failure of maturational pruning (larger caudate volumes were found in recent ADHD studies Castellanos et al 1994, Mataro et al 1997.

Most volumetric studies of the striatum in schizophrenia have found enlargement of the striatal region in patients with schizophrenia Breier et al 1992, Buchanan et al 1993, Heckers et al 1991, Hokama et al 1995, Jernigan et al 1991, Swayze and Yuh 1992 and evidence so far suggests that this enlargement is a consequence of neuroleptic treatment. Studies by Chakos et al 1994, Keshavan et al 1994, Elkashef et al 1994, and Rodriguez and colleagues (1997) noted that the increase in caudate volume in schizophrenic patients followed treatment with typical neuroleptics. Following the introduction of atypical neuroleptics, several follow-up studies noted a decrease in volume when patients were switched from typical to atypical neuroleptics (Chakos et al 1995, Frazier et al 1996, Westmoreland et al 1997, Westmoreland Corson 1999, in press). However, all of these studies either examined patients who were neuroleptic naive at baseline (and then following their exposure to medication for a few months and assessing volume changes over time) or measured striatal volumes of chronically ill patients who had been medicated with neuroleptics for many years, and these did not take into account striatal volumes of patients versus those of control subjects at baseline.

Our study (comparing neuroleptic naive-schizophrenic patients with matched control subjects) was instituted to address the question as to whether or not there is an independent underlying structural abnormality of the striatum potentially attributable to the disease process, which precedes volume changes induced by neuroleptic medication. Of the few direct comparisons between caudate volumes of first-episode neuroleptic-naive schizophrenic patients and control subjects, two preliminary reports found the caudate nucleus to be significantly smaller in seven never-medicated schizophrenic patients versus eighteen control subjects (Shihabuddin et al 1998) and sixteen never-medicated patients versus seventeen control subjects (Keshavan et al 1998). In the present report, we address this question using a larger sample of neuroleptic naive patients and control subjects and highly reliable automated measurement techniques. We postulated that if indeed the caudate volumes of never-medicated schizophrenic patients and control subjects were found to be significantly different, this might point to factors in the disease pathology as being additionally responsible for aberrations in caudate volume in patients versus control subjects. If volumes were not found to be significantly different, this would strengthen the argument that neuroleptic treatment may be solely responsible for the structural changes in the striatal region of neuroleptic medicated patients.