THE JOURNAL OF COMPARATIVE NEUROLOGY 323:326-340 (1992)
Sylvian Fissure Morphology and Asymmetry in Men and Women: Bilateral Differences in Relation to Handedness in Men SANDRA F. WITELSON AND DEBRA L. KIGAR Departments of Psychiatry (D.L.K.) and Biomedical Sciences (S.F.W.,D.L.K.), McMaster University, Hamilton, Ontario L8N 3%5Canada
ABSTRACT The anatomy of the sylvian fissure in the human brain was studied to develop reliable criteria for anatomical landmarks of the posterior part of the fissure for use in its definition and measurement; to quantify right-left asymmetries in segments o f t he sylvian fissure; to assess whether any anatomical features are associated with hand preference (selected as one index of hemispheric functional asymmetry); and whether structure-function relationships are similar in men and women. A sample of 67 brain specimens (24 men and 43 women, mean age = 53 years) was studied postmortem (with the aid of dissection) from people who had been tested before death for detailed hand preference. Sylvian fissure anatomy in the human brain is very variable and no agreement exists as to the point of its posterior termination. The posterior ascending ramus, origmating at the posterior bifurcation of the fissure, was found to be the continuation of the main limb of the sylvian fissure. Three segments of the sylvian fissure were defined and measured: anterior, horizontal, and vertical. The anterior segment showed no asymmetry; the horizontal segment was twice as large on the left side as on the right; and the vertical segment twice as large on the right. The two asymmetries counterbalanced each other, and overall asymmetry in the posterior region (horizontal plus vertical) was minimal. The basic asymmetry is in the position at which the fissure turns up, resulting in the different extent and position of the surrounding right and left parietal and temporal gyri and associated cytoarchitectonic regions. The possible embryological course of the asymmetry is discussed. Handedness correlated with anatomy of the sylvian fissure in men. In contrast to general expectation, hand preference was associated with a bilateral feature of morphology, and not with less asymmetry in non-right-handers. Men having consistent-right-hand preference had longer horizontal segments in both hemispheres compared to men not having consistentright-hand preference. The direction and magnitude of asymmetry did not differ between the two male hand-preference groups. Since hand preference is art index of other motor and perceptual functions which are asymmetrically represented in the two hemispheres in gyri surrounding the sylvian fissure, it is suggested that anatomy of the sylvian fissure is related to functional asymmetries in men. A sex difference in structure-function relationship was observed. No association was found between hand preference and sylvian fissure anatomy in women. These results corroborate a similar sex difference in the structure-function relationship observed for the corpus callosum (Witelson; Bruin 112:799-835, 1989). They are also consistent with neuropsychological findings of sex differences in the localization of cognitive functions in parietotemporal regions, such as differences in degree of functional asymmetry and in the pattern of intrahemispheric representation of cognitive functions. It is suggested that there is a greater dissociation between motoric and perceptual asymmetries in women than in men. IYYL WiIey-I,ihs, lnc Key words: autopsy, cerebral cortex, cerebral dominance, embryology, neuropsychological tests, sex characteristics
~
Accepted May 7, 1992. Address reprint requests to Dr. S.F. Witelson, Department of Psychiatry, McMaster University, 1200 Main St. West, Hamilton, Ont. IAN 32!5, Canada.
O 1992 WILEY-LISS, INC.
SYLVIAN FISSURE ANATOMY, HANDEDNESS, AND SEX The lateral or sylvian fissure (SF) is one of the primary fissures of the primate brain; unlike the other primary fissures, the SF is not formed by an infolding of the cortex, but results from uneven growth of the outer cortex relative to inner structures (Cunningham, 1892). The corpus striatum and its covering cortex, the insula, do not grow as quickly as the surrounding neocortex and consequently the SF is formed by the opercularization of neocortex over the insula (Cunningham, 1892). Unique to the human brain is the extensive degree of opercularization of frontal, parietal, and temporal lobe regions which form the banks of the SF. This anatomical feature allows for great morphological variation among hemispheres, both between and within brains. SF variability was noted early on and became the focus of several extensive studies of the human brain as early as a century ago (e.g., Eberstaller, 1890; Cunningham, 1892; Shellshear, '37). However, there is still no agreement on the anatomical definition of the SF and, accordingly, various issues concerning its delineation and quantitation remain to be resolved. Since many temporal and parietal gyri surrounding the SF depend for their definition on SF anatomy, delineation of the SF is essential for quantitative studies of adjacent cortical regions. The SF is surrounded by gyri which mediate higher perceptual and motoric functions. Many of these cognitive functions are asymmetrically represented in the right and left hemispheres in the human brain. This aspect of brain organization, referred to as hemispheric functional asymmetry or functional asymmetry, is most marked in the human brain. The localization of functions in the gyri surrounding the SF has been documented most clearly by study of the cognitive deficits in patients with unilateral, localized brain damage. For example, in most people, the posterior part of the superior temporal gyrus on the left side is crucial for the comprehension of spoken language; on the right side, this region is involved in attention; on the left side, the supramarginal gyrus which surrounds the end of the SF is involved in the processing of aspects of motoric behavior (praxis);on the right side, the parietal operculum mediates aspects of spatial cognition such as orientation in three-dimensional space (Hecaen and Albert, '78; Heilman and Valenstein, '85). Because various functional asymmetries are not perfectly correlated with each other (Bryden et al., '83), it is appropriate to refer to functional asymmetry in the plural. The pattern of functional asymmetries varies in direction and magnitude. For example, left handers have a greater prevalence of right-hemisphericrepresentation of language function than do right handers (HQcaenet al., '81). Numerous studies indicate that the magnitude of functional asymmetry for language functions is less in left than right handers (Bryden, '82) and in women than men (Beaton, '85).
Abbreviations
ASF CLH CRH H-type H&V-type HSF non-CRH PAR PDR SF V-type VSF
anterior segment of SF consistent-left-hander(handed) consistent-right-hander (handed) horizontal type of SF horizontal and vertical SF type horizontal segment of SF non consistent-right-hander posterior ascending ramus posterior descending ramus sylvian fissure vertical type of SF vertical segment of SF
327
Fig. 1. Tracing of the lateral aspect of a left hemisphere in an adult human brain showing the general morphology of the SF and associated landmarks. AB is the main horizontal branch of SF. Several rami (labelled 1to 5) may appear to be connected to the SF, but on dissection are seen to be separate sulci within the surrounding gyri. A, point at which the anterior ascending ramus (AAR)and anterior horizontal ramus (AHR) join the SF; C and PC, dorsal ends of the central and postcentral sulci; C1 and PC1, points where the central and postcentral sulci meet or are extended to meet SF; H, point at which the main transverse sulcus of Heschl on the supratemporal plane meets or can be extended to the lateral edge of the SF; B, point of bifurcation of the posterior end of the main horizontal branch of SF; S, anterior end of the SF, defined as the most anterior point on the lateral aspect where the frontal and temporal lobes do or can touch each other; S1, end of the posterior ascending ramus (PAR) and the posterior limit of SF; Sz, end of the posterior descending ramus (PDR); STG, superior temporal gyrus; STS, superior temporal sulcus; 1, diagonal sulcus; 2, sulcus subcentralis anterior; 3, sulcus subcentralis posterior; 4,sulcus retrocentralis transversus; 5, sulcus supratemporalis transversus posterior, respectively, all based on terminology from Bailey and von Bonin ('51). The full extent of the SF and its rami are shown by the bold line.
The SF and the morphology of the surrounding gyri are inextricably associated with each other, and in this respect the SF may be a marker of variation in anatomy and function of opercular regions. Figure 1 shows the general morphology of the SF in the adult human brain. It includes a main horizontal branch (often called the posterior horizontal limb or stem), which extends from the frontal through the parietal and temporal lobes (AB, Fig. 1).The SF usually branches into rami at the anterior and posterior ends. There are several superficial sulci in the gyri surrounding the main horizontal branch which appear to be rami of the SF, but these are not connected to the SF (see 1to 5, Fig. 1). The anatomy of the SF is sufficiently variable and complex that it is not always obvious from inspection of the lateral surface which sulci are rami of the SF at its termination. Accordingly, many studies of the SF were descriptive in nature, and, when measurements were made, they focused on the central portion of the S F which is more reliably delineated. Although many of the early studies examined large samples, statistical analyses of the results were not given. Eberstaller (18901, as described in Cunningham (18921, was among the first to study the SF and noted the difficulty in delineating both anterior and posterior SF limits. In an attempt to select reliable landmarks Eberstaller chose to measure only the main horizontal branch (AB, Fig. 1) and reported that, in 170 left and 183right hemispheres, mean length was greater on the left side by 6.5 mm and that this asymmetry occurred in 63% of specimens. After Eberstall-
328 er's (1890) report, most authors adopted his anatomical criteria and used the posterior branching point (B) as the end of the SF. Eberstaller's findings were substantiated in several similar studies. For example, Cunningham (1892) measured the SF, defined as the distance between points S and B (Fig. l),and found a longer left than right fissure in a sample of 23 left and 28 right hemispheres. Cunningham (1892) further noted that the angulation of the SB segment of SF relative to the axis of the brain's maximal anteroposterior length was asymmetrical (by 4"), such that the right SF sloped upward more steeply than the left and that the divergence in slope occurred mainly in the postcentral regions. Subsequent researchers focused on the postcentral horizontal segment (CIB,Fig. 1)of the main horizontal branch and documented that the left postcentral segment was longer than the right (Shellshear, '37; Connolly, '50; Rubens et al., '76). Connolly ('50) examined the external morphology of the posterior terminating rami of the SF (BS1, BS2, Fig. 1) in a sample of 60 brains. He did not measure them but suggested that since BS1, the posterior ascending ramus (PAR), appears to be the more stable ramus than the posterior descending ramus (PDR), PAR is likely the main posterior branch of SF. Rubens et al. ('76) were among the first to measure a SF segment using the end of PAR as the end of the SF. They measured the distance CISl (Fig. 1)in a group of 36 brains and found that it was longer on the left side in 75% of cases, but did not reach statistical significance. Their measure was the shortest distance between points C1 and S1and did not reflect the curved course of SF. Yeni-Komshian and Benson ('76) reported a statistically significantly longer left than right SF in a series of 25 human brains in which the end of SF was defined as either PAR or PDR depending on which appeared from external examination to be a deeper sulcus. Studies of S F anatomy are reviewed in greater detail elsewhere (Geschwind, '74; Rubens, '77; Witelson, '77; Geschwind and Galaburda, '85; Witelson and Kigar, '88). Although there has been considerable study of the anatomy of the SF, there is little information concerning whether variation in SF anatomy is correlated with variation in behavior or functional asymmetry. A century ago, Cunningham (1892, pp. 135, 136) anticipated this issue. Noting the more extensive and inferiorly positioned left SF, he questioned: "Can it be due to a greater development on the left side of that part of the cortex in which the motor centres reside? Had it only been evident in man, I should have been inclined to associate it with right-handedness." Cunningham had noted similar but less marked anatomic asymmetry in nonhuman apes, but at the time of his work, the phenomenon of cerebral dominance (that is, functional asymmetry) was considered to be dependent on language and handedness and therefore unique to people (Witelson, '87a). The current general conception of functional asymmetries includes two main types of processing, each predominantly represented in one hemisphere, and is more general and applicable to preverbal infants and nonhuman animals who also show functional asymmetries (Witelson, '87b). Some evidence is available that neuroanatomical asymmetries assessed by various in vivo imaging methods are associated with functional asymmetries. LeMay and Culebras ('72) noted that asymmetry in SF anatomy may be inferred from arteriographs which show the angulation of the middle cerebral artery as it emerges from the SF and loops over the parietal operculum. They reported that right
S.F. WITELSON AND D.L. KIGAR handers showed less arterial angulation in the left than right hemisphere and that left handers showed less asymmetry. Less asymmetry in occipital lobe breadth based on computed tomography (CT) scans was found in left than right handers (LeMay, '77). Using direct measurement of postmortem specimens, midsagittal area of the corpus callosum, particularly area of the isthmus (the posterior part of the callosal body), was found to be larger in non-right-handers than right handers (Witelson, '85, '89; Witelson and Goldsmith, '91). The isthmus includes interhemispheric axons from posterior parietal and temporal regions surrounding the SF (reviewed in Witelson, '89). It was suggested that the larger callosal areas of non-righthanders is a basis of their greater bihemispheric representation of cognitive functions. A detailed review of the association between neuroanatomical features and functional asymmetries is given elsewhere (Witelson and Kigar, '88). These results suggest that variation in morphological features of the SF may relate to measures of variation in functional asymmetries. For the isthmus, there is a marked sex difference in structure-function relationship: the association between isthmal area and hand preference was observed in men, but not women (Witelson, '89). Similar interactions between the factors of hand preference and sex in regard to callosal anatomy have since been reported in magnetic resonance imaging (MRI) studies using a similar classification of hand preference (e.g., Habib et al., '91). Such results suggest that variation in the anatomy of the posterior SF may be associated also with sex. The overall aim of the present study was to assess the relationship between variation in the anatomy of the posterior region of the SF and one aspect of functional asymmetry in men and women. The specific aims of the study were to develop reliable criteria to define gross anatomical landmhrks of the posterior SF of the adult human brain; to measure various segments of the SF to assess its asymmetry; to assess SF morphology in relation to measures of hand preference used as an index of functional asymmetry; and to assess whether any structurefunction relationships for the SF are different between the sexes.
MATERIALS AND METHODS Subject source The subjects were cancer patients who were recruited over the past 13 years (as of April 1990) for a study which required them to take a battery of neuropsychological tests and to consent to a postmortem examination, with autopsy permission to be granted by their next-of-kin at the time of death. Each subject gave written informed consent for both aspects of the project which had the approval of the institutional ethics review board. All subjects were aware that they had metastatic disease which was amenable only to palliative radiotherapy or chemotherapy, and that their prognosis for long-term survival was poor. The subjects were ambulatory and free of neurological symptoms or any debilitating symptoms when they started the project. The detailed procedures used for recruitment, testing, follow-up of subjects, and autopsy request are presented elsewhere (Witelson and McCulloch, '91). This report is based on brain specimens obtained from 71 consecutively obtained autopsies from a total group of 127 research subjects who had been recruited to the project, of
'
SYLVIAN FISSURE ANATOMY, HANDEDNESS, AND SEX whom 113 had died. The sample of 71 cases included 25 men and 46 women, varying in age from 25 to 70 years. The difference in number between the sexes is primarily due to the sex difference in the prevalence of the main types of cancer (lung and breast) of the patients involved in the project. Further details such as the percent of cases with different types of malignancies and the number of years of survival subsequent to entry into the project are given elsewhere (Witelson and McCulloch, ’91).
Hand preference test A modified version of Annett’s (’67) hand questionnaire was used, in which the subject was asked to demonstrate hand use, as opposed to just reporting hand preference, on a series of 12 unimanual (e.g., tooth brushing) and bimanual (e.g., threading a needle) tasks. For each item, preference was recorded as right, either, or left. Two scoring systems were used: a categorical (dichotomous) system for hand preference, and a continuous measure of handedness. Several different classification systems of hand preference are typically used in psychological studies: for example, either right- or left-handed defined by writing hand; or right-, mixed- or left-handed, defined by arbitrary criteria of the percentage of tasks done with the right and left hands. The classification used in this study is based on Annett’s (’85) model of the classification and genetics of hand preference. Two categories were used. Right preference was defined stringently as consistentright-hand preference (CRH) (all “right” or no “left” preferences). The other category consisted of those who were not CRH, that is, non consistent-right-hand preference (non-CRH) (defined as left-hand preference on at least 1 of the 12 tasks, regardless of which hand is used for writing). The non-CRH category includes people who show mixed-hand preference (MH) (defined as any combination of right and left preferences) and consistent-left-hand preference (CLH) (defined as all “left” or no “right” preferences). Of the 71 brain specimens, two subjects did not have sufficient handedness testing required for unequivocal classification. This reduced the sample to 69 cases. The prevalence of CRH, MH, and CLH in the sample of 69 cases was 62, 36, and 1% (there was one CLH), respectively, which is similar to that of 66, 30, and 4% observed in large samples (Annett, ’72). A history of hand preference was also obtained to determine whether any subjects were forced to change their writing hand. Three of the men were forced dextrals. In each case, they preferred their left hand for other tasks and therefore were classified as non-CRH, regardless of writing hand. Each of the 12 items recorded as right, either, or left preference, were scored as +1, 0, or -1, respectively. Therefore, the minimumimaximum possible scores were - 121+ 12, which provided a continuous measure of handedness from 100%left- to 100%right-hand preference.
Brain removal procedure The autopsies were done in six local hospitals following a standard procedure established for the study. The entire brain including the medulla was removed by the pathologist during autopsy. On removal from the skull, the fresh brain was weighed, and immediately immersed in a solution of 10%neutral buffered formalin. The brain was suspended by the basilar artery in a large bucket lined with cotton batting to prevent distortion of any surfaces. The mean time
329
interval between death and brain fixation a t autopsy was 10.9 hours, minimax = 1122 hours (n = 57 cases). The remaining 12 brains were obtained by a different administrative route in which the person bequeathed hislher body to science, and this involved a different fixation procedure. The whole body was perfused with embalming fixative with additional perfusion of the brain via the internal carotid arteries. The brain was removed 16 days after death, in compliance with Ontario law, and then immersed in 10% formalin.
Brain weights Fresh brain weight at autopsy was available for the 57 brains obtained from autopsy. Upon arrival in the laboratory, each brain was weighed again within 2 days after death. The brain was then kept in a refrigerated storage room with formalin changed after 1 week, 1 month, and thereafter every 12 months. A third brain weight was obtained 3 weeks after death. The 12 bequeathal-obtained brains were weighed in the laboratory 3 days after immersion fixation (that is, 19 days after death). All laboratory weights were obtained with a Sartorius balance (model U3600), accurate to 0.1 g. Initial increase in brain weight after death is reported to occur within the first 12 hours (Appel and Appel, ’42). Thereafter insignificant changes in weight and volume are reported with formalin fixation [approximately 2% in either direction up to a t least 7 weeks after death (Small and Peterson, ’8211. We found similar results in our series. For the brains obtained from autopsy, mean fresh weight and mean 2-day weight were 1,310 g and 1,342 g, respectively. This is a 2.4%difference (t = 4.0, df = 55, P < 0.001). The mean 3-week weight was 1,327 g which is 1.1% lower than the 2-day weight (t = 3.9, df = 55, P < 0.001). The 3-week weight for the autopsy-obtained brains was designed to be roughly comparable to the 3-day immersion weight (3 days after immersion fixation but approximately 3 weeks after death) for the bequeathal-obtained brains. Weights of the cerebral hemispheres for all brain specimens were obtained after dissection as described in the following section.
Anatomical dissection After the brain was well fixed and weight measurements obtained, photography of the whole brain was done for each specimen. The cerebellum was then removed and the brainstem cut off at the level just below the inferior colliculus. The brain was bisected in the midsagittal plane and the meninges and blood vessels were removed from all surfaces of the cerebral hemispheres. Each hemisphere was weighed and the lateral aspect was photographed for SF measurement. The maximal fronto-occipital length of each hemisphere was measured with a millimeter ruler placed on the flat midsagittal brain surface. Hemisphere length was used as a baseline measure in the analysis of SF measures. Each hemisphere was then dissected to reveal the floor of the full SF as described in a later section. Neuropathological examination of gross features was performed at autopsy and a t each dissection stage, with particular attention paid to atrophy, hemorrhages, and metastases or malformations. Histopathological assessment of tissue was done for any suspect tissue. Of the 69 brains having hand-preference classification, two were excluded due to neuropathology affecting the gross anatomy of SF regions. One woman had polymicrogyria in the left superior temporal gyrus and left parietal operculum, a
S.F. WITELSON AND D.L. KIGAR
330
TABLE 1. Mean and Standard Deviation (in parentheses) for Baseline Measures of Age, Brain Weight, and Handedness Scores Brain weight (el ~~~
Submoup
re a t death Abn years
No
Men CRH Non-CRH Total Women CRH Non-CRH Total
13 115
24
29 14 43
67
Total ~~~
HemisphereL
Whole brain'
RightJ
Left
Handedness score4
53 '12) 56 (11) 541111
1,431 (85) 1,440 (126) 1,435 I1031
595 138) 590 157) 592 (461
592 \31l 592 (56, 592 1431
11.2 11.11 2.8 (8.2) 7 4 16.91
62 (9) 52 (9)
1,275 (91) 1,209 187) 1,254 (94)
524 ( 3 5 ) 513 146)
62 I91
521 (38)
5'24 139, 507 1391 519 1401
11.6 11.2) 6.5 15.2) 9.9 ' 3 91
53 t l 0 l
1,318l130)
547 (53)
545 I541
9.0 15.31
~
Summaiy of two-way ANOVAs
Sylvian Fissure Morphology and Asymmetry in Men and Women: Bilateral Differences in Relation to Handedness in Men SANDRA F. WITELSON AND DEBRA L. KIGAR Departments of Psychiatry (D.L.K.) and Biomedical Sciences (S.F.W.,D.L.K.), McMaster University, Hamilton, Ontario L8N 3%5Canada
ABSTRACT The anatomy of the sylvian fissure in the human brain was studied to develop reliable criteria for anatomical landmarks of the posterior part of the fissure for use in its definition and measurement; to quantify right-left asymmetries in segments o f t he sylvian fissure; to assess whether any anatomical features are associated with hand preference (selected as one index of hemispheric functional asymmetry); and whether structure-function relationships are similar in men and women. A sample of 67 brain specimens (24 men and 43 women, mean age = 53 years) was studied postmortem (with the aid of dissection) from people who had been tested before death for detailed hand preference. Sylvian fissure anatomy in the human brain is very variable and no agreement exists as to the point of its posterior termination. The posterior ascending ramus, origmating at the posterior bifurcation of the fissure, was found to be the continuation of the main limb of the sylvian fissure. Three segments of the sylvian fissure were defined and measured: anterior, horizontal, and vertical. The anterior segment showed no asymmetry; the horizontal segment was twice as large on the left side as on the right; and the vertical segment twice as large on the right. The two asymmetries counterbalanced each other, and overall asymmetry in the posterior region (horizontal plus vertical) was minimal. The basic asymmetry is in the position at which the fissure turns up, resulting in the different extent and position of the surrounding right and left parietal and temporal gyri and associated cytoarchitectonic regions. The possible embryological course of the asymmetry is discussed. Handedness correlated with anatomy of the sylvian fissure in men. In contrast to general expectation, hand preference was associated with a bilateral feature of morphology, and not with less asymmetry in non-right-handers. Men having consistent-right-hand preference had longer horizontal segments in both hemispheres compared to men not having consistentright-hand preference. The direction and magnitude of asymmetry did not differ between the two male hand-preference groups. Since hand preference is art index of other motor and perceptual functions which are asymmetrically represented in the two hemispheres in gyri surrounding the sylvian fissure, it is suggested that anatomy of the sylvian fissure is related to functional asymmetries in men. A sex difference in structure-function relationship was observed. No association was found between hand preference and sylvian fissure anatomy in women. These results corroborate a similar sex difference in the structure-function relationship observed for the corpus callosum (Witelson; Bruin 112:799-835, 1989). They are also consistent with neuropsychological findings of sex differences in the localization of cognitive functions in parietotemporal regions, such as differences in degree of functional asymmetry and in the pattern of intrahemispheric representation of cognitive functions. It is suggested that there is a greater dissociation between motoric and perceptual asymmetries in women than in men. IYYL WiIey-I,ihs, lnc Key words: autopsy, cerebral cortex, cerebral dominance, embryology, neuropsychological tests, sex characteristics
~
Accepted May 7, 1992. Address reprint requests to Dr. S.F. Witelson, Department of Psychiatry, McMaster University, 1200 Main St. West, Hamilton, Ont. IAN 32!5, Canada.
O 1992 WILEY-LISS, INC.
SYLVIAN FISSURE ANATOMY, HANDEDNESS, AND SEX The lateral or sylvian fissure (SF) is one of the primary fissures of the primate brain; unlike the other primary fissures, the SF is not formed by an infolding of the cortex, but results from uneven growth of the outer cortex relative to inner structures (Cunningham, 1892). The corpus striatum and its covering cortex, the insula, do not grow as quickly as the surrounding neocortex and consequently the SF is formed by the opercularization of neocortex over the insula (Cunningham, 1892). Unique to the human brain is the extensive degree of opercularization of frontal, parietal, and temporal lobe regions which form the banks of the SF. This anatomical feature allows for great morphological variation among hemispheres, both between and within brains. SF variability was noted early on and became the focus of several extensive studies of the human brain as early as a century ago (e.g., Eberstaller, 1890; Cunningham, 1892; Shellshear, '37). However, there is still no agreement on the anatomical definition of the SF and, accordingly, various issues concerning its delineation and quantitation remain to be resolved. Since many temporal and parietal gyri surrounding the SF depend for their definition on SF anatomy, delineation of the SF is essential for quantitative studies of adjacent cortical regions. The SF is surrounded by gyri which mediate higher perceptual and motoric functions. Many of these cognitive functions are asymmetrically represented in the right and left hemispheres in the human brain. This aspect of brain organization, referred to as hemispheric functional asymmetry or functional asymmetry, is most marked in the human brain. The localization of functions in the gyri surrounding the SF has been documented most clearly by study of the cognitive deficits in patients with unilateral, localized brain damage. For example, in most people, the posterior part of the superior temporal gyrus on the left side is crucial for the comprehension of spoken language; on the right side, this region is involved in attention; on the left side, the supramarginal gyrus which surrounds the end of the SF is involved in the processing of aspects of motoric behavior (praxis);on the right side, the parietal operculum mediates aspects of spatial cognition such as orientation in three-dimensional space (Hecaen and Albert, '78; Heilman and Valenstein, '85). Because various functional asymmetries are not perfectly correlated with each other (Bryden et al., '83), it is appropriate to refer to functional asymmetry in the plural. The pattern of functional asymmetries varies in direction and magnitude. For example, left handers have a greater prevalence of right-hemisphericrepresentation of language function than do right handers (HQcaenet al., '81). Numerous studies indicate that the magnitude of functional asymmetry for language functions is less in left than right handers (Bryden, '82) and in women than men (Beaton, '85).
Abbreviations
ASF CLH CRH H-type H&V-type HSF non-CRH PAR PDR SF V-type VSF
anterior segment of SF consistent-left-hander(handed) consistent-right-hander (handed) horizontal type of SF horizontal and vertical SF type horizontal segment of SF non consistent-right-hander posterior ascending ramus posterior descending ramus sylvian fissure vertical type of SF vertical segment of SF
327
Fig. 1. Tracing of the lateral aspect of a left hemisphere in an adult human brain showing the general morphology of the SF and associated landmarks. AB is the main horizontal branch of SF. Several rami (labelled 1to 5) may appear to be connected to the SF, but on dissection are seen to be separate sulci within the surrounding gyri. A, point at which the anterior ascending ramus (AAR)and anterior horizontal ramus (AHR) join the SF; C and PC, dorsal ends of the central and postcentral sulci; C1 and PC1, points where the central and postcentral sulci meet or are extended to meet SF; H, point at which the main transverse sulcus of Heschl on the supratemporal plane meets or can be extended to the lateral edge of the SF; B, point of bifurcation of the posterior end of the main horizontal branch of SF; S, anterior end of the SF, defined as the most anterior point on the lateral aspect where the frontal and temporal lobes do or can touch each other; S1, end of the posterior ascending ramus (PAR) and the posterior limit of SF; Sz, end of the posterior descending ramus (PDR); STG, superior temporal gyrus; STS, superior temporal sulcus; 1, diagonal sulcus; 2, sulcus subcentralis anterior; 3, sulcus subcentralis posterior; 4,sulcus retrocentralis transversus; 5, sulcus supratemporalis transversus posterior, respectively, all based on terminology from Bailey and von Bonin ('51). The full extent of the SF and its rami are shown by the bold line.
The SF and the morphology of the surrounding gyri are inextricably associated with each other, and in this respect the SF may be a marker of variation in anatomy and function of opercular regions. Figure 1 shows the general morphology of the SF in the adult human brain. It includes a main horizontal branch (often called the posterior horizontal limb or stem), which extends from the frontal through the parietal and temporal lobes (AB, Fig. 1).The SF usually branches into rami at the anterior and posterior ends. There are several superficial sulci in the gyri surrounding the main horizontal branch which appear to be rami of the SF, but these are not connected to the SF (see 1to 5, Fig. 1). The anatomy of the SF is sufficiently variable and complex that it is not always obvious from inspection of the lateral surface which sulci are rami of the SF at its termination. Accordingly, many studies of the SF were descriptive in nature, and, when measurements were made, they focused on the central portion of the S F which is more reliably delineated. Although many of the early studies examined large samples, statistical analyses of the results were not given. Eberstaller (18901, as described in Cunningham (18921, was among the first to study the SF and noted the difficulty in delineating both anterior and posterior SF limits. In an attempt to select reliable landmarks Eberstaller chose to measure only the main horizontal branch (AB, Fig. 1) and reported that, in 170 left and 183right hemispheres, mean length was greater on the left side by 6.5 mm and that this asymmetry occurred in 63% of specimens. After Eberstall-
328 er's (1890) report, most authors adopted his anatomical criteria and used the posterior branching point (B) as the end of the SF. Eberstaller's findings were substantiated in several similar studies. For example, Cunningham (1892) measured the SF, defined as the distance between points S and B (Fig. l),and found a longer left than right fissure in a sample of 23 left and 28 right hemispheres. Cunningham (1892) further noted that the angulation of the SB segment of SF relative to the axis of the brain's maximal anteroposterior length was asymmetrical (by 4"), such that the right SF sloped upward more steeply than the left and that the divergence in slope occurred mainly in the postcentral regions. Subsequent researchers focused on the postcentral horizontal segment (CIB,Fig. 1)of the main horizontal branch and documented that the left postcentral segment was longer than the right (Shellshear, '37; Connolly, '50; Rubens et al., '76). Connolly ('50) examined the external morphology of the posterior terminating rami of the SF (BS1, BS2, Fig. 1) in a sample of 60 brains. He did not measure them but suggested that since BS1, the posterior ascending ramus (PAR), appears to be the more stable ramus than the posterior descending ramus (PDR), PAR is likely the main posterior branch of SF. Rubens et al. ('76) were among the first to measure a SF segment using the end of PAR as the end of the SF. They measured the distance CISl (Fig. 1)in a group of 36 brains and found that it was longer on the left side in 75% of cases, but did not reach statistical significance. Their measure was the shortest distance between points C1 and S1and did not reflect the curved course of SF. Yeni-Komshian and Benson ('76) reported a statistically significantly longer left than right SF in a series of 25 human brains in which the end of SF was defined as either PAR or PDR depending on which appeared from external examination to be a deeper sulcus. Studies of S F anatomy are reviewed in greater detail elsewhere (Geschwind, '74; Rubens, '77; Witelson, '77; Geschwind and Galaburda, '85; Witelson and Kigar, '88). Although there has been considerable study of the anatomy of the SF, there is little information concerning whether variation in SF anatomy is correlated with variation in behavior or functional asymmetry. A century ago, Cunningham (1892, pp. 135, 136) anticipated this issue. Noting the more extensive and inferiorly positioned left SF, he questioned: "Can it be due to a greater development on the left side of that part of the cortex in which the motor centres reside? Had it only been evident in man, I should have been inclined to associate it with right-handedness." Cunningham had noted similar but less marked anatomic asymmetry in nonhuman apes, but at the time of his work, the phenomenon of cerebral dominance (that is, functional asymmetry) was considered to be dependent on language and handedness and therefore unique to people (Witelson, '87a). The current general conception of functional asymmetries includes two main types of processing, each predominantly represented in one hemisphere, and is more general and applicable to preverbal infants and nonhuman animals who also show functional asymmetries (Witelson, '87b). Some evidence is available that neuroanatomical asymmetries assessed by various in vivo imaging methods are associated with functional asymmetries. LeMay and Culebras ('72) noted that asymmetry in SF anatomy may be inferred from arteriographs which show the angulation of the middle cerebral artery as it emerges from the SF and loops over the parietal operculum. They reported that right
S.F. WITELSON AND D.L. KIGAR handers showed less arterial angulation in the left than right hemisphere and that left handers showed less asymmetry. Less asymmetry in occipital lobe breadth based on computed tomography (CT) scans was found in left than right handers (LeMay, '77). Using direct measurement of postmortem specimens, midsagittal area of the corpus callosum, particularly area of the isthmus (the posterior part of the callosal body), was found to be larger in non-right-handers than right handers (Witelson, '85, '89; Witelson and Goldsmith, '91). The isthmus includes interhemispheric axons from posterior parietal and temporal regions surrounding the SF (reviewed in Witelson, '89). It was suggested that the larger callosal areas of non-righthanders is a basis of their greater bihemispheric representation of cognitive functions. A detailed review of the association between neuroanatomical features and functional asymmetries is given elsewhere (Witelson and Kigar, '88). These results suggest that variation in morphological features of the SF may relate to measures of variation in functional asymmetries. For the isthmus, there is a marked sex difference in structure-function relationship: the association between isthmal area and hand preference was observed in men, but not women (Witelson, '89). Similar interactions between the factors of hand preference and sex in regard to callosal anatomy have since been reported in magnetic resonance imaging (MRI) studies using a similar classification of hand preference (e.g., Habib et al., '91). Such results suggest that variation in the anatomy of the posterior SF may be associated also with sex. The overall aim of the present study was to assess the relationship between variation in the anatomy of the posterior region of the SF and one aspect of functional asymmetry in men and women. The specific aims of the study were to develop reliable criteria to define gross anatomical landmhrks of the posterior SF of the adult human brain; to measure various segments of the SF to assess its asymmetry; to assess SF morphology in relation to measures of hand preference used as an index of functional asymmetry; and to assess whether any structurefunction relationships for the SF are different between the sexes.
MATERIALS AND METHODS Subject source The subjects were cancer patients who were recruited over the past 13 years (as of April 1990) for a study which required them to take a battery of neuropsychological tests and to consent to a postmortem examination, with autopsy permission to be granted by their next-of-kin at the time of death. Each subject gave written informed consent for both aspects of the project which had the approval of the institutional ethics review board. All subjects were aware that they had metastatic disease which was amenable only to palliative radiotherapy or chemotherapy, and that their prognosis for long-term survival was poor. The subjects were ambulatory and free of neurological symptoms or any debilitating symptoms when they started the project. The detailed procedures used for recruitment, testing, follow-up of subjects, and autopsy request are presented elsewhere (Witelson and McCulloch, '91). This report is based on brain specimens obtained from 71 consecutively obtained autopsies from a total group of 127 research subjects who had been recruited to the project, of
'
SYLVIAN FISSURE ANATOMY, HANDEDNESS, AND SEX whom 113 had died. The sample of 71 cases included 25 men and 46 women, varying in age from 25 to 70 years. The difference in number between the sexes is primarily due to the sex difference in the prevalence of the main types of cancer (lung and breast) of the patients involved in the project. Further details such as the percent of cases with different types of malignancies and the number of years of survival subsequent to entry into the project are given elsewhere (Witelson and McCulloch, ’91).
Hand preference test A modified version of Annett’s (’67) hand questionnaire was used, in which the subject was asked to demonstrate hand use, as opposed to just reporting hand preference, on a series of 12 unimanual (e.g., tooth brushing) and bimanual (e.g., threading a needle) tasks. For each item, preference was recorded as right, either, or left. Two scoring systems were used: a categorical (dichotomous) system for hand preference, and a continuous measure of handedness. Several different classification systems of hand preference are typically used in psychological studies: for example, either right- or left-handed defined by writing hand; or right-, mixed- or left-handed, defined by arbitrary criteria of the percentage of tasks done with the right and left hands. The classification used in this study is based on Annett’s (’85) model of the classification and genetics of hand preference. Two categories were used. Right preference was defined stringently as consistentright-hand preference (CRH) (all “right” or no “left” preferences). The other category consisted of those who were not CRH, that is, non consistent-right-hand preference (non-CRH) (defined as left-hand preference on at least 1 of the 12 tasks, regardless of which hand is used for writing). The non-CRH category includes people who show mixed-hand preference (MH) (defined as any combination of right and left preferences) and consistent-left-hand preference (CLH) (defined as all “left” or no “right” preferences). Of the 71 brain specimens, two subjects did not have sufficient handedness testing required for unequivocal classification. This reduced the sample to 69 cases. The prevalence of CRH, MH, and CLH in the sample of 69 cases was 62, 36, and 1% (there was one CLH), respectively, which is similar to that of 66, 30, and 4% observed in large samples (Annett, ’72). A history of hand preference was also obtained to determine whether any subjects were forced to change their writing hand. Three of the men were forced dextrals. In each case, they preferred their left hand for other tasks and therefore were classified as non-CRH, regardless of writing hand. Each of the 12 items recorded as right, either, or left preference, were scored as +1, 0, or -1, respectively. Therefore, the minimumimaximum possible scores were - 121+ 12, which provided a continuous measure of handedness from 100%left- to 100%right-hand preference.
Brain removal procedure The autopsies were done in six local hospitals following a standard procedure established for the study. The entire brain including the medulla was removed by the pathologist during autopsy. On removal from the skull, the fresh brain was weighed, and immediately immersed in a solution of 10%neutral buffered formalin. The brain was suspended by the basilar artery in a large bucket lined with cotton batting to prevent distortion of any surfaces. The mean time
329
interval between death and brain fixation a t autopsy was 10.9 hours, minimax = 1122 hours (n = 57 cases). The remaining 12 brains were obtained by a different administrative route in which the person bequeathed hislher body to science, and this involved a different fixation procedure. The whole body was perfused with embalming fixative with additional perfusion of the brain via the internal carotid arteries. The brain was removed 16 days after death, in compliance with Ontario law, and then immersed in 10% formalin.
Brain weights Fresh brain weight at autopsy was available for the 57 brains obtained from autopsy. Upon arrival in the laboratory, each brain was weighed again within 2 days after death. The brain was then kept in a refrigerated storage room with formalin changed after 1 week, 1 month, and thereafter every 12 months. A third brain weight was obtained 3 weeks after death. The 12 bequeathal-obtained brains were weighed in the laboratory 3 days after immersion fixation (that is, 19 days after death). All laboratory weights were obtained with a Sartorius balance (model U3600), accurate to 0.1 g. Initial increase in brain weight after death is reported to occur within the first 12 hours (Appel and Appel, ’42). Thereafter insignificant changes in weight and volume are reported with formalin fixation [approximately 2% in either direction up to a t least 7 weeks after death (Small and Peterson, ’8211. We found similar results in our series. For the brains obtained from autopsy, mean fresh weight and mean 2-day weight were 1,310 g and 1,342 g, respectively. This is a 2.4%difference (t = 4.0, df = 55, P < 0.001). The mean 3-week weight was 1,327 g which is 1.1% lower than the 2-day weight (t = 3.9, df = 55, P < 0.001). The 3-week weight for the autopsy-obtained brains was designed to be roughly comparable to the 3-day immersion weight (3 days after immersion fixation but approximately 3 weeks after death) for the bequeathal-obtained brains. Weights of the cerebral hemispheres for all brain specimens were obtained after dissection as described in the following section.
Anatomical dissection After the brain was well fixed and weight measurements obtained, photography of the whole brain was done for each specimen. The cerebellum was then removed and the brainstem cut off at the level just below the inferior colliculus. The brain was bisected in the midsagittal plane and the meninges and blood vessels were removed from all surfaces of the cerebral hemispheres. Each hemisphere was weighed and the lateral aspect was photographed for SF measurement. The maximal fronto-occipital length of each hemisphere was measured with a millimeter ruler placed on the flat midsagittal brain surface. Hemisphere length was used as a baseline measure in the analysis of SF measures. Each hemisphere was then dissected to reveal the floor of the full SF as described in a later section. Neuropathological examination of gross features was performed at autopsy and a t each dissection stage, with particular attention paid to atrophy, hemorrhages, and metastases or malformations. Histopathological assessment of tissue was done for any suspect tissue. Of the 69 brains having hand-preference classification, two were excluded due to neuropathology affecting the gross anatomy of SF regions. One woman had polymicrogyria in the left superior temporal gyrus and left parietal operculum, a
S.F. WITELSON AND D.L. KIGAR
330
TABLE 1. Mean and Standard Deviation (in parentheses) for Baseline Measures of Age, Brain Weight, and Handedness Scores Brain weight (el ~~~
Submoup
re a t death Abn years
No
Men CRH Non-CRH Total Women CRH Non-CRH Total
13 115
24
29 14 43
67
Total ~~~
HemisphereL
Whole brain'
RightJ
Left
Handedness score4
53 '12) 56 (11) 541111
1,431 (85) 1,440 (126) 1,435 I1031
595 138) 590 157) 592 (461
592 \31l 592 (56, 592 1431
11.2 11.11 2.8 (8.2) 7 4 16.91
62 (9) 52 (9)
1,275 (91) 1,209 187) 1,254 (94)
524 ( 3 5 ) 513 146)
62 I91
521 (38)
5'24 139, 507 1391 519 1401
11.6 11.2) 6.5 15.2) 9.9 ' 3 91
53 t l 0 l
1,318l130)
547 (53)
545 I541
9.0 15.31
~
Summaiy of two-way ANOVAs
