Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
, 104 (7), 2513-8

Brain Shape in Human Microcephalics and Homo Floresiensis


Brain Shape in Human Microcephalics and Homo Floresiensis

Dean Falk et al. Proc Natl Acad Sci U S A.


Because the cranial capacity of LB1 (Homo floresiensis) is only 417 cm(3), some workers propose that it represents a microcephalic Homo sapiens rather than a new species. This hypothesis is difficult to assess, however, without a clear understanding of how brain shape of microcephalics compares with that of normal humans. We compare three-dimensional computed tomographic reconstructions of the internal braincases (virtual endocasts that reproduce details of external brain morphology, including cranial capacities and shape) from a sample of 9 microcephalic humans and 10 normal humans. Discriminant and canonical analyses are used to identify two variables that classify normal and microcephalic humans with 100% success. The classification functions classify the virtual endocast from LB1 with normal humans rather than microcephalics. On the other hand, our classification functions classify a pathological H. sapiens specimen that, like LB1, represents an approximately 3-foot-tall adult female and an adult Basuto microcephalic woman that is alleged to have an endocast similar to LB1's with the microcephalic humans. Although microcephaly is genetically and clinically variable, virtual endocasts from our highly heterogeneous sample share similarities in protruding and proportionately large cerebella and relatively narrow, flattened orbital surfaces compared with normal humans. These findings have relevance for hypotheses regarding the genetic substrates of hominin brain evolution and may have medical diagnostic value. Despite LB1's having brain shape features that sort it with normal humans rather than microcephalics, other shape features and its small brain size are consistent with its assignment to a separate species.

Conflict of interest statement

The authors declare no conflict of interest.


Fig. 1.
Fig. 1.
Comparisons of right lateral views of virtual endocasts from 10 normal humans (Upper, blue) and 9 microcephalics (Lower, blue). Discriminant and canonical analyses classify the virtual endocast of LB1 [417,f (Upper, red)] with normal humans and those from a female human dwarf (752,f, Lower) and the Basuto woman (358,f, Lower) with microcephalics. Images are labeled with their cranial capacities and sex: f, female; m, male (see Table 1 for details about individual specimens).
Fig. 2.
Fig. 2.
Key for right lateral (a), posterior (b), and inferior (c) views of endocasts. Measurements (3, a chord; others projected): (a) 1, cerebral length (fp-op); 2, cerebellar pole-projected frontal pole; 3, anterior cerebral height (chord v-tp); 4, temporal pole-projected frontal pole; (b) 5, cerebral width (right and left points that define maximum projected width); 6, cerebellar width (right and left most lateral points on cerebellum, includes sigmoid sinus if visible in posterior view); 7, endocast height (v-b); (c) 8, frontal breadth (mat-mat). Landmarks: b, midpoint on line tangent to the base of cerebellum; bs, intersection of right brainstem with right temporal lobe; cp, most caudal point on the cerebellum in lateral view (may be on either side); fp, most rostral point on the frontal lobes in lateral view (may be on either side); mat, in basal view, point at lateral edge of endocast located at level of the most anterior point of the temporal lobe; op, most caudal point on occipital lobes in lateral view (may be on either side); tp, most rostral point on temporal lobes in lateral view (may be on either side); v, vertex. Four indices used in discriminant and canonical analyses: cerebellar protrusion = 2/1; relative length posterior base = (–4)/1; relative cerebellar width = 6/5; relative frontal breadth = 8/6.
Fig. 3.
Fig. 3.
Scatter plot of relative frontal breadth on cerebellar protrusion. The legend of Fig. 2 contains a description of the measurements that were used to create the ratios. Discriminant analysis demonstrated that these two variables classified microcephalics (M) and normal humans (N) with 100% success. The dwarf, Basuto woman (BW), and LB1, which were not used to develop the classification functions, were classified, respectively, as two microcephalics and a normal human.

Similar articles

See all similar articles

Cited by 12 PubMed Central articles

See all "Cited by" articles

Publication types

LinkOut - more resources