Our findings in the brainstem are similar to previous results of other authors. Pedhazur EJ (1982) Multiple regression in behavioral research.

The total intracranial volume, TICV, was treated as a confounder to correct for differences in physiognomy. Using a 1.5 T unit (Magnetom Vision, Siemens, Erlangen, Germany) with standard head coil, two MRI sequences were acquired.

Subsequently, the model described by formula (2) was fitted (Fig. This area is a good estimate of pontine volume assuming that the pons is rotation-symmetrical (Raz et al., 1992). Braitenberg and Atwood identified the regions with large amounts of folial white matter: the anterior vermal region and a sagittal band in the mid-hemisphere (Braitenberg and Atwood, 1958). Two raters, blinded for the other one's results, measured cerebellar and brainstem volume in 15 subjects. This interpretation is supported by regional volumetry of the cerebral hemispheres. The ‘medial hemisphere' was defined as one-quarter of the total width of the hemisphere. The hemispheres should behave like the posterior vermis, but the contrary was observed.

(c,d) Region-growing-based segmentation identified the full delineation of brainstem and cerebellum. In: Robbins pathologic basis of disease (Cotran RS, Kumar V, Robbins SL, eds), pp.

The linear regression model for total cerebellar volume (TCV) included age and the total intracranial volume (TICV) as independent variables (Table 2). None of the subjects met DSM-IIIR criteria for alcohol abuse. The curve predicted that the volume remained stable until age 50 years and declined thereafter. Women had larger TICV-corrected volumes mH2 and V2 than men. TICV did not meet the inclusion criterion of Punivariate = 0.1 for regions mH2, V1, V2 and V3. This shift also excluded the red nucleus and part of the substantia nigra from the brainstem volume (Fig. This is suggested by similarities between the patterns of atrophy and compartments defined by the distribution of genes and proteins. None of the volunteers reported to drink or have drunken more than socially accepted.

), pp. The posterior boundary was marked by a coronal plane through the posterior commissure and the obex. Amsterdam: Elsevier. Considering the recent advances in the knowledge of functional as well as morphological compartmentalization of the cerebellum, we believe that a precise characterization of the regional patterns of cerebellar and brainstem age-related atrophy is necessary. Kleinbaum DG, Kupper LL, Muller KE, Nizam A (1998) Applied regression and multivariate methods .

Evidence for the latter comes from studies demonstrating connections between the lateral cerebellum/dentate nucleus and parietal or frontal association areas in the contralateral cerebral cortex (Crosby, 1969; Middleton and Strick, 1994). Most authors correct for body height (Raz et al., 1997) or TICV (Escalona et al., 1991; Schaefer et al., 1991; Shah et al., 1991; Doraiswamy et al., 1992).

Raz N (1996) Neuroanatomy of the aging brain observed in vivo. Accordingly, the lateral hemisphere region would be highly exposed to vascular atrophy, which is not consistent with our findings. The cortex covers the outer portion (1.5mm to 5mm) of … neurons, glia cells, afferent/efferent axons, dendrites — see Discussion) degenerates, the volume should converge at value b.

In the cerebellum, watershed areas were identified in the deep white matter (which may be protected by high collateralization), the lateral hemispheres (SCA–PICA watershed), and the petrosal surface (SCA–AICA–PICA watersheds) (Gillilan, 1969; Savoiardo et al., 1987; Cormier et al., 1992). Histology and computerized tomography studies reported pronounced volume loss in the anterior vermis (Koller et al., 1981; Torvik et al., 1986). Subsequently, automated region-growing segmentation was applied (Fig. Humans: In Comparison with the Ventral Tegmental Area/Substantia Nigra Pars The magnitude of age effect on different regions was compared by testing two partial correlation coefficients for significant difference.

The TICV included brain and CSF volumes caudally delimited by the foramen magnum. Other possible explanations for discrepant results in cerebellar volumetry are racial or social differences in the subject samples of different studies (N. Raz, personal communication). The quadratic model better described the dependency between age and total cerebellar volume (partial F-test: P < 0.05). This may partly account for the higher shrinkage in the anterior vermis, but cannot be the only factor, because the hemisphere shows only minor changes with age. The cerebellar peduncles which were part of the middle brainstem region could not be consistently separated using landmarks. To demonstrate effects of age only, other independent covariates are omitted in these plots. The partial correlation coefficient of the lateral hemisphere was compared to the coefficients of the other regions (Table 4). Since this test, as well as tests for significant correlation, were performed multiple times, Bonferroni-corrected significance levels were used. For young age groups, the quadratic model predicted an increase in volume between age 20 and 30 years, which seemed unlikely to be real. When Bonferroni correction was not used, all vermal coefficients and the coefficients for mH2 and mH3 were significantly different from the lateral hemisphere coefficient (Table 4). The pattern of volume loss observed in the present study is similar to alcoholic cerebellar atrophy (Victor et al., 1959; Torvik and Torp, 1986; Karhunen et al., 1994).

Among the radial segments, the caudoventral segments (V3 and mH3­) were larger than the superior and posterior segments.