No abstract

Although three-dimensional electron microscopy (3D-EM) permits structural characterization of macromolecular assemblies in distinct functional states, the inability to classify projections from structurally heterogeneous samples has severely limited its application. We present a maximum likelihood-based classification method that does not depend on prior knowledge about the structural variability, and demonstrate its effectiveness for two macromolecular assemblies with different types of conformational variability: the Escherichia coli ribosome and Simian virus 40 (SV40) large T-antigen.

Algebraic reconstruction techniques (ART) are iterative procedures for recovering objects from their projections. It is claimed that by a careful adjustment of the order in which the collected data are accessed during the reconstruction procedure and of the so-called relaxation parameters that are to be chosen in an algebraic reconstruction technique, ART can produce high-quality reconstructions with excellent computational efficiency. This is demonstrated by an example based on a particular (but realistic) medical imaging task, showing that ART can match the performance of the standard expectation-maximization approach for maximizing likelihood (from the point of view of that particular medical task), but at an order of magnitude less computational cost.

A prospective sample of 69 healthy adults, age range 18-80 years, was studied with magnetic resonance imaging scans (T2 weighted, 5 mm thick) of the entire cranium. Volumes were obtained by a segmentation algorithm that uses proton density and T2 pixel values to correct field inhomogeneities ("shading"). Average (-SD) brain volume, excluding cerebellum, was 1090.91 ml (±# 114.30; range, 822.19-1363.66), and cerebrospinal fluid (CSF) volume was 127.91 ml (±57.62; range, 34.00-297.02). Brain volume was higher (by 5 ml) in the right hemisphere (P < 0.0001). Men (n = 34) had 91 ml higher brain and 20 ml higher CSF volume than women (n = 35). Age was negatively correlated with brain volume [r(67) =-0.32, P < 0.01] and positively correlated with CSF volume (r = 0.74, P < 0.0001). The slope of the regression line with age for CSF was steeper for men than women (P = 0.03). This difference in slopes was significant for sulcal (P < 0.0001), but not ventricular, CSF. The greatest amount of atrophy in elderly men was in the left hemisphere, whereas in women age effects were symmetric. The findings may point to neuroanatomic substrates of hemispheric specialization and gender differences in age-related changes in brain function. They suggest that women are less vulnerable to age-related changes in mental abilities, whereas men are particularly susceptible to aging effects on left hemispheric functions.The study of brain regulation of human behavior requires measurement of structural variables, and this has been done primarily by postmortem studies (e.g., refs. 1-6). Atrophy was inferred from reduced brain weight or volume or increased differences between brain volume and cranial capacity-i.e., cerebrospinal fluid (CSF) volume. Several studies found aging associated with atrophy (7-9). Others did not find age effects until senescence (usually defined as age >55 or 60; refs. 4, 10, and 11). Women have lower brain volume, related to body and cranial size (e.g., refs. 4, 6, and 9).In vivo measurement of brain volume became feasible with computed tomography, and more recently with magnetic resonance imaging (MRI), which is more sensitive than computed tomography for determining sulcal changes and has better tissue contrast, multiplanar imaging capabilities, absence of bone artifact, and no ionizing radiation. In addition to elucidating structural substrates of brain function, anatomic volume measures are important for interpreting metabolic data (12). Thus, decline in cerebral blood flow and metabolism with age (e.g., refs. 13-15), and higher cerebral blood flow in women (16), could be explained by structural effects (17).Several computed tomography studies investigated ageassociated changes. Takeda and Matsuzawa (18) subjects aged 21-81 studied with a 2-T magnet and a spinlattice relaxation time (T1)-weighted sequence. However, in contrast to the other studies, which yielded brain volume estimates averaging 1100-1200 ml, their estimates were >2000 ml for men and >1800 ml for women. Jernigan et al. (23) found a linear re...