Elsevier

Neuropsychologia

Volume 38, Issue 10, 1 September 2000, Pages 1375-1381
Neuropsychologia

Comparison of overall brain volume and midsagittal corpus callosum surface area as obtained from NMR scans and direct anatomical measures: a within-subject study on autopsy brains

https://doi.org/10.1016/S0028-3932(00)00048-8Get rights and content

Abstract

Formalin fixed brains obtained from autopsy were studied by magnetic resonance imaging, using FLASH and MPR sequences. The overall volume and the midsagittal surface area of the corpus callosum (cc) of these brains were also directly measured by water displacement for overall volume determination and morphometry for cc surface area measures, allowing comparisons of these measures in a within-subject design. Thin NMR sections (1 mm thickness for FLASH and 1.25 mm for MPR) yielded overall brain volume estimates that did not differ significantly from direct volume estimations but thicker (5 mm) sections led to a significant overestimation of brain volume for NMR measures relative to direct displacement measures. There was no overall effect of type of NMR measure: FLASH and MPR scans did not yield significantly different values for overall brain volume or callosal surface area. Direct and NMR measures of the corpus callosum surface at midsagittal level did not yield significantly different estimates.

Introduction

Even a cursory review of the brain size literature reveals that there are large differences in the estimate of ‘average brain size’ in Homo sapiens. Much of the variability relates to sampling procedures, but there is also a substantial discrepancy between studies that examine relatively similar samples [19]. Overall brain size has numerous important theoretical implications, both in the evolutionary context [1], [14] and in the context of models of brain function. For example, it is not irrelevant for models of brain function whether the cortical surface area is the generally accepted average figure of 2400–2500 cm2 with some 100 billion estimated neurons [18], [22] or whether the surface area and neuron count is very much less than that [8], [15]. For humans, the canonical value of brain weight in the literature appears to be in the range of 1400–1450 g but numerous studies yield far lower weights [19].

This note reports an attempt to test the comparability of brain volume and areal estimates based on NMR imaging with those based on post mortem anatomy. It thus follows a time-honored tradition of attempts to correlate brain volume or weight measures for different methods. One of the earliest such studies was done by Davis [9] who used cranial measures obtained by the Viennese physician Weisbach to calculate estimated brain weights, which were then compared with the fresh brain weights obtained by Weisbach. In our study, the object was to see whether the NMR-derived values correlate well enough with autopsy measures to encourage full use of the potential advantages offered by NMR technology. In obtaining brain measures, NMR procedures offer better experimental control than is available for autopsy samples, because subject age, sex, health status, body parameters, mental status and any other variable of interest can be selected deliberately, and a priori. In addition, NMR studies offer the opportunity to conduct within- subject longitudinal studies. However, NMR-derived values of brain volume or other measures depend on a reconstruction of numerical data. Thus, the validity of the final outcome depends on the algorithms employed, which differ across different laboratories. It comes as no surprise that rather different volumes are given for brain volume by different researchers [3], [4], [7], [10], [19], [20], [21], and the differences may be due to different methods of reconstructing NMR images from the numerical data. In the above studies there is no independent way to validate the accuracy of brain volume measures because only within subject design using different methods can achieve this. The present study offers an independent validation of NMR-derived measures by comparing autopsy brains which were analyzed both with traditional morphometric or displacement methods with NMR scans and reconstructions. In our study, we sidestep the problems that would have arisen if a reverse order of methods had been used, that is, taking NMR scans prior to death and comparing these to post mortem material. While that would also have been a within-subject design, there can be changes in brain volume at death, notably through edema [2], [23], which make comparisons between NMR and autopsy measures problematic.

Section snippets

Subjects

Ten brains (eight males and two females, average age 80.3 years, SD=5 years) were studied. For the purposes of this study, the advanced age of the subjects is not likely to be of importance.

Direct volume determination

Meninges, and major blood vessels were removed. The brains were turned and shaken for about 30 s to eliminate some of the adhering water, with lateral tilting of the brain and care was taken to allow as much ventricular water as possible to drain via the openings provided by the removal of the pituitary

Results and discussion

Table 1 shows the various estimates of weight and volume. It can be seen that there is a very considerable reduction in substance once the meninges are stripped and the CSF is taken into account, and that the brain volumes are relatively low. The most relevant series, and the one which is demographically similar to the one from which the present set of brains was drawn from is the one described by Zilles [24], and the low brain volumes are in keeping with the volumes obtained for individuals

Conclusion

This study was prompted by the question as to how direct brain measures correlate with NMR-derived estimates. For the purposes of general studies, in which large numbers of brains are compared, the NMR based estimates appear adequate, with the proviso that such estimates are based on thin rather than thick sections. For thick sections, the overestimate of brain tissue volume is significant both in statistical terms and in terms of the effect size. However, while the correspondences between

Acknowledgments

The authors would like to thank Kai Lutz and Christoph Klein for their help in data preparation. The first author was supported by NSERC Grant A 7054, the second author was supported by DFG JA 737/4-1 and JA 737/5-1 and the third author supported by DFG KZ (SFB194/A6).

References (24)

  • E. Bigler et al.

    Aging, brain size and IQ

    Intelligence

    (1995)
  • K.M. Bishop et al.

    Sex differences in the human corpus callosum

    Neuroscience and Biobehavioral Reviews

    (1997)
  • M. Peters et al.

    Unsolved problems in comparing brain sizes in homo sapiens

    Brain and Cognition

    (1998)
  • F. Aboitiz

    Does bigger mean better? Evolutionary determinants of brain size and structure

    Brain, Behaviour, and Evolution

    (1996)
  • F.W. Appel et al.

    Intracranial variation in the weight of the human brain

    Human Biology

    (1942)
  • A.J. Bartley et al.

    Genetic variability of human brain size and cortical gyral patterns

    Brain

    (1997)
  • N. Bushmakin

    Characteristics of the brain of the mongol race

    American Journal of Physical Anthropology

    (1928)
  • V.S. Caviness et al.

    The human brain age 7–11 years: a volumetric analysis based on magnetic resonance images

    Cerebral Cortex

    (1996)
  • C. Cherniak

    The bounded brain: toward quantitative neuroanatomy

    Journal of Cognitive Neuroscience

    (1990)
  • J.B. Davis

    Contributions towards determining the weight of the brain in different races of man

    Philosophical Transactions, Series B (London)

    (1868)
  • J.N. Giedd et al.

    Quantitative magnetic resonance imaging of human brain development: ages 4–18

    Cerebral Cortex

    (1996)
  • A. Hrdlicka

    Physical changes in human and other brains collected under different conditions and preserved in various formalin preparations

    Proceedings US National Museum

    (1906)
  • Cited by (21)

    • Head size may modify the impact of white matter lesions on dementia

      2012, Neurobiology of Aging
      Citation Excerpt :

      However, there are possible measurement errors on CT. Beam-hardening artifacts and partial volume-averaging may obscure the precise inner boundary of the bone and the outer boundary of the brain, and slice thickness and separation of slices add to these errors (Peters et al., 2000). However, if anything, these errors should decrease the possibilities of finding differences between the groups.

    View all citing articles on Scopus
    View full text