Quantification of frontal and temporal lobe brain-imaging findings in schizophrenia: a meta-analysis

• Published in: Psychiatry research 2003-02, Vol.122 (2), p.69-87
• Main Authors: Davidson, Lara L., Heinrichs, R.Walter
• Format: Article
• Language: English
• Subjects: Adult; Adult and adolescent clinical studies; Biological and medical sciences; Case-Control Studies; Effect sizes; Energy Metabolism - physiology; Female; Frontal Lobe - pathology; Frontal Lobe - physiopathology; Humans; Magnetic Resonance Imaging; Male; Medical sciences; Middle Aged; MRI; PET; Psychology. Psychoanalysis. Psychiatry; Psychopathology. Psychiatry; Psychoses; Quantitative review; Reference Values; Regional Blood Flow - physiology; Schizophrenia; Schizophrenia - diagnosis; Schizophrenia - physiopathology; Sensitivity and Specificity; Temporal Lobe - pathology; Temporal Lobe - physiopathology; Tomography, Emission-Computed
• ISSN: 0925-4927; 0165-1781; 1872-7506
• DOI: 10.1016/S0925-4927(02)00118-X

Magnetic resonance imaging (MRI) and positron emission tomography (PET) studies of the frontal and temporal lobes in schizophrenia patients and healthy controls have proliferated over the past 2 decades, but there have been relatively few attempts to quantify the evidence. In this meta-analytic review, 155 studies on frontal and temporal lobe neurobiology were synthesized, reflecting results from 4043 schizophrenia patients and 3977 normal controls. Cohen's d was used to quantify case-control differences, and moderator variable analysis indexed the relation of sample and imaging characteristics to the magnitude of these differences. Frontal metabolic and blood flow deficiencies in conjunction with cognitive activation tasks (‘hypofrontality’) emerged as the strongest body of evidence, demonstrating abnormalities that distinguish approximately half of schizophrenia patients from healthy people. Most case-control comparisons with structural and functional imaging yield small and in many cases unstable findings. Technical scanning parameters like slice thickness and magnet strength did not vary with case-control differences consistently across the meta-analyses. However, patient sample characteristics including sample size, handedness and gender composition emerged frequently as moderators of brain-imaging effect sizes.