MSU LIBRARIES ._;—_ RETURNING MATERIALS: P1ace in book droB_to remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. 2 / 9"“- @3929 1 f INTERHEHISPHERIC RELATIONSHIPS BETWEEN HWOTOPICAL CmTICAL REGIONS: SEX AND HANDEDNESS DIFFERENCES IN HU‘IAN REGIONAL CEREBRAL BLOOD FLOW By Steven Uarach A DISSERTATION Submitted to Hichigm State [hiversity in 'partial fulfillment of the requirements for the degree of mCTOR G PHILOSOPHY Department of Psychology and Newoacienee Progra 1983 ABSTRACT INTERHEHISPHERIC RELATIONSHIPS BETWEEN HCMOTOPICAL CORTICAL REGIONS: SEX AND HANDEDNESS DIFFERENCES IN HUIAN REGIONAL CEREBRAL BLOOD FLW By Steven Harach 0n the basis of behavioral data. it has been proposed that the cerebral hemispheres of males are more functionally asymmetric than those of females, and that the hemispheres of right-handers are more functionally asymmetric than those of left-handers. his study examined functional asymmetries using a physiological measu'e of brain function: regional cerebral blood flow (rCBF). The subjects were 15 right-handed males, Wright-handed females, 15 left-handed males. and 17 left-handed females. Each subject was stuliied with the 133-xenon inhalation technique. Measuruents of rCBF were made at 8 pairs of homotopical loci du'ing each of three cognitive conditions: resting, solving verbal analogies, and solving spatial [tabla-s. Right-left asymmetries in the absolute manituie of blood flow indicated asymmetrical activity. Right-left partial correlations (controlling for' total brain blood flow) between homotopical loci addressed the issue of functional interplay between the hemispheres, more positive correlations indicating a more symmetrical (excitatory) interaction. A greater nunber of positive interhemispheric correlations and more strongly positive correlations were observed for females than males and for right-handers than left-handers. Regional asymmetries were more frequent for males than for females and more fr'equent among right-handers than left-handers. These sex and handedness differences were more evident du'ing rest than duing cognitive activity. The sex differences were most evident at the middle and superior precentral regions, and the handedness differences were most noticable in the superior precentral and superior postcentral regions. Differences were not found at all homotopical pairs of loci. In fact, at some loci group differences opposite to the general pattern were observed. The results support the hypothesis that male brains are' more functionally asymmetric than female brains with the following qualification: group differences in interhemispheric functioning are not ubiquitous but seem to have a regional and task specificity. Data relevant to the hypothesis of'greater asymmetry in right-handers appear contradictory.‘ The correlational results contradict the hypothesis, whereas the asymmetry data support it. The apparent contradiction may reflect the greater heterogeneity of brain organization among left-handers. Capyright by STEVEN JAY HARACH 1983 ii ACKNOWLEDGMENTS I thank the following people for their contributions to this dissertation: Lynwood Clemens, Snlomo Eliash, Ruben Gar, Raquel Gar, Lauren Harris, John Johnson, Antonio Nunez, Martin Reivich, and Susan Resnick. This project was supported by USPHS grant NS-10939—09, a Spencer Foundation grant (to Ruben C. Our), and NIH grant NH 30u56. - The Department of Psychology and the Neuroscience Program of Michigan State University generously supported me throughout my graduate studies. A Graduate Fellowship from the National Science Foundation enabled me to carry out this work. I am especially gratefhl to John Johnson, my major professor and advisor since my undergraduate days, and to Ruben Oar, with mom I worked closely on this study. 111 TABLE OF CONTENTS LIST OF TABLES LIST OF F IGLHES IN TRODUC TI ON INDIRECT MEASURES SEX DIFFERENCES HANDEDNESS DIFFERENCES DIRECT MEASURES INTERHEHISPHERIC CORRELATIONS AND ASYMMETRIES METHOD SUBJECTS PROCEDURE THE 133-XENON INHALATION TECHNIQUE STATISTICAL ANALYSIS RESULTS - SEX DIFFERENCES HA NDE DNESS DIFFERENCES DISCUSSION SEX DIFFERENCES HANDEDNESS DIFFERENCES GENERAL DISCLBSION LIST OF REFERENCES iv PAGE TABLE TABLE TABLE TABLE TABLE TABLE TABLE TABLE 7. 8. LIST (1“ TABLES SEX DIFFERENCES IN RIGHT-LEFT CORRELATIONS HANDEDNESS DIFFERENCES IN RIGHT-LEFT CORRELATIONS SEX DIFFERENCES IN RIGHT-LEFT PARTIAL CORRELATIONS HANDEDNESS DIFFERENCES IN RIGHT-LEFT PARTIAL CORRELATIONS SEX DIFFERENCES IN RIGHT-LEFT ASTHMETRIES SEX DIFFERENCES IN RIGHT-LEFT ASYHMETRIES FOR HANDEDNESS GROUPS HANDEDNESS DIFFERENCES IN RIGHT-LEFT ASYHHETRIES HANDEDNESS DIFFERENCES IN RIGHT-LEFT ASYMMETRIES FOR SEX GROUPS PAGE 57 59 60 6 1 62 63 61‘ FIGURE FIGURE FIGURE FIGURE FIGURE FIGURE FIGURE FIGURE FIGURE 7. 8. 9. FIGURE 10. FIGURE 11. FIGURE 12. LIST OF FIGURES . EXAMPLE OF 133-XE CLEARANCE CURVE LOCATION a" DETECTORS SEX DIFFERENCES IN CORRELATIONS DURING REST SEX DIFFERENCES IN PARTIAL CORRELATIONS DURING REST SEX DIFFERENCES IN PARTIAL CORRELATIONS DURING REST: HANDEDNESS GROUPS SEX DIFFERENCES IN ASYMMETRY DURING REST SEX DIFFERENCES IN PARTIAL CORRELATIONS SEX DIFFERENCES IN ASYMMETRIES HANDEDNESS DIFFERENCES IN PARTIAL CORRELATIONS HANDEDNESS DIFFERENCES IN PARTIAL CORRELATIONS DURING REST: SEX GROUPS HANDEDNESS DIFFERENCES IN ASYMMETRY DURING REST HANDEDNESS DIFFERENCES IN ASYMMETRIES vi PA GE 33 3‘3 39 an 43 us an ‘49 51 53 55 I NTRODUC TION Knowledge of human brain function is primarily inferential. For many questions of interest technical and ethical considerations have maie impossible the use of most methods that measu'e nenrophysiology _i_g 1312 for many questions of interest. The study of non-hula brains, observations of presuptive behavioral indices of brain function, and natu'al ”experiments" in neu'ology and neu'osu'gery have been the nearly exclusive sonrces of ideas on how the normal human brain works. These indirect approaches entail certain risks. H'ominent among them is the necessity of making assumptions about the relationship between the indirect measu'e and the normal neu'ophysiology it pu'ports to, describe. These assunptions are inherently difficult to validate. since validation requires making the measu'e 33.222 that could not be obtained in the first place. A body of literatu'e based unon these indirect measures inevitably contains few conclusions for which any consensus has been reached. The study of lateral asymmetries in the hunun brain is a case in point. The winciple that in nearly all right-handers the left hemisphere plays the leading role in language functions, the right-hemisphere in non-verbal, \risuospatial. and certain notional» functions is well-established (e.g.. Dimond and Beamont, 19718). What is less clear -- and the subject of some dispute at present -- is the extent to unich this modal pattern may be correlated with the sex and handedness of the subject. INDIRECT MEASURES To illustrate some of the problems in using the indirect methods, an ex-ination of the two most caamon techniques in nenropsycmlogy follows. koca's description of a lesion in the inferior frontal cortex of an aphasic patient (Race, 1861) established the major paradigm of nenropsychology. Studies of behavioral abnormalities consequent to neu'al lesions, incu'red either by accident or surgical necessity, have povided insight into cerebral localization not only of sensory-motor but also higher cognitive functions (e.g.. Daria, 1966). The accnracy with which a brain lesion can be identified and localized solely on the basis of clinical interview and behavioral testing (Luria, 1966: Golden et al.. 1980) is a testament to the validity of this paradigm. However, this predictive validity is not perfect (e.g.. 711! correct localization for the Daria-Nebraska Neuopsychological httery: Cohen et al., 1981), and for much work such checks on the conclusions do not exist. The most notable modern example of this paradigm is the investigation of split-brain patients. ‘nne work of Sperry and his collaborators has daonstrated that verbal abilities are preferentially the domain of the left hemisphere and spatial unilities are [referentially right hemisphere functions (e.g.. mzzaniga. Hogan, 3. Sperry. 1965:' Gazzaniga& Sperry. 1967: Sperry, 19711). However, the generalizability of conclusions reached "on these studies is not clear, since the brains of commissu'otomy patients are not normal to begin with. This surgery is performed on patients with a long history of uncontrollable epilepsy that is generalized or multifocal in origin. Profound intellectual deficits are often resent in these patients. ‘me working physiological assunnption of this lesion strategy is that any behavior that is changed following dame to the brain is a normal ‘nanction of the tissue inju‘ed. Implicit in this assunnption is that the neu'ophysiological effects" of daage are restricted to the site of lesion: in the case of commissu'otomy patients. the hemispheres are assumed to be functionally unchanged, merely disconnected. me to the successful application of this technique and the lack of technical alternatives, this assnmnption has rarely been questioned, bu. it is undountedly u'ong. lunged neuons degenerate. This phenomenon is so well established that it not only can be found in neu-oanatomical textbooks (e.g.. B'odal. 1%1) but has served as the basis for techniques to trace axonal connections between brain regions. Degeneration distal to an injnred non is inevitable, proximal to it likely, and trananeu-onal to it caamon. For exannple, Ebner and Myers (1965) duonstrated widespread bilateral degeneration in the neocortex following cutting of the corpus callosnmn and anterior commissure in cats and racoons, and Glickstein and mitteridge (1976) observed degeneration of Layer III pyraidal cells in Area 18 of cats after the cutting of the corpus callosnmn or destruction of the homotopical cortex. Thus degeneration caused by mental daaqe is not restricted locally but is ’present at sites distant from the lesion site. Functional changes subsequent to nenral dame are also expected. These incluie collateral sprouting from remaining input fibers to the denervated cells (e.g.. Tripp and Hells, 1978). neu'otranamitter receptor supersensitivity in denervated cells (e.g.. Crease et al.. 1977). and changes in the physiological response properties of the denervated cells (e.g.. Millar et al.. 1976). A'lesioned brain is not a normal brain with one piece missing: it is abnormal away from the lesion site as well as within it. A second paradigm to assess functional asymmetry involves normal sunjecta stuiied ndth inatrnmnents, such as the tachistoscope, that permit lateralized wesentation of cognitive tasks. In tachistoscopic resentation. sunjects are required to fix their gaze to the center of the visual field. Stimuli. such as a vertical crangmnent of a three-letter nonsense syllwle, are flashed to one side of the fixation point for a du'ation too brief to allow saccades. The wocedue is repeated for a set of a given stimulus type, and the wesentation varies randannly between right and left visual fields. If the subject correctly identifies a greater nunber of stimuli in one visual field than the other, the experimenter concluies that the cerebral hemisphere contralateral to that visual field (the initial hemisphere to receive the visual stimulus) excels in the woeessing of stimuli of the type presented. This conclusion is considered valid if callosal tranaaission occnrs at the sane speed from right to left'as them left to right. The absence of a visual field bias would imply that the hemispheres are equipotential for the task or that a ronghly equivalent bilateral involvement is necessary. The length of time the stimulus is exposed to the subject is critical. The longer the exposu'e time. the longer the opportunity for interhemispheric camunication and. therefore, the less noticeable the effects of a truly lateralized function. The exposue time is either identical for each subject (e.g.. (raves et al.. 1981) or determined individually for each snbject (e.g.. Levy and Reid, 1976). It is usually a fixed annount added to the briefest exposnre time at which the subject can correctly solve a simplified version of the task. Studying interindividual or intergreu: differences tachistoscopically makes the further assumptions that the speed of callosal transmission and the nuaber of synapses required for processing are constant from sunject to subject. If. for exaaple. callosal tranaaission were quicker in females than males, a tachistoscopic stuiy using a constant exposu'e time criterion would show a smaller visual field bias in females even if their brains had identical functional asymmetry. This condition, in fact, may be the case. The corpus callesunn of females has been shown to have a thicker spleniua, that portion connecting visual and other posterior cortices, than that of males (Delacoste-Utansing and Holloway. 1982a). If this size difference is due to the [resence of thicker fibers, thicker myelination, or a greater woportion of thick or myelinated fibers, faster callosal transaission would be expected for females. The hypothesized sex difference in functional asymmetry, as induced them tachistoscopic studies, might be an artifact of a sex difference in the speed of callosal tranaaission. Given the limitations of these. indirect measures of n-eu'ophysiology, it is not surp‘ising that a point would be reached in neu'opsychology at which uniformity of opinion was noticeably missing. SEX DIFFERENCES Considerable evidence has accrued to support the hypothesis that male brains are more asymmetrically organized than female brains, yet many reject this interpretation (of. McGlone, and anmmentary, 1%0). I-Nidence that verbal skills are characterized by a greater bilateral involvennent in females includes observations that left hemisphere lesions are associated with more severe speech distu'bances (McGlone. 1977) and agreater decreaent in verbal relative to performance IQ (McGlene. 1978: Inglis and Dawson. 1982: Inglis et al.. 1982) annong males than females. stimulation of left inferior frontal cortex is followed by a greater incidence of manning errors in male versus female patients (1hteer et al.. 1982). and alternate injection of sodium aaytal into each hemisphere leads to a greater discrepancy for males in oral fluncy (lelone. 1982). Right hemisphere lesions are reported to cause less severe impairment of spatial functions in females (Hecaen et al.. 1981). implying greater bilateral involvmnnent 'in spatial functions. lbwever. other studies (e.g.. Lansdell. 1%8: McGlone and Kertesz. 1973) found no significant interaction of sex and hemisphere in the cognitive disturbance subsequent to lesions. Snch clinical studies have been criticized for the small sample sizes used. lack of control for the extent,of lesion. and for other methodological problems (Kinsbonrne. 1980: McGlene. 1980). The interpretation of greater female symmetry in verbal functions has. also been challenged as possibly due to the females' panorbid superiority in verbal functions (Sherman. 1980). Dichotic listening studies of normal subjects have been taken as support for the idea that males show a greater lateralization of verbal Mctions (Lake and anyden. 1976: Springer and Searleaan. 1918). These studies found a greater right ear advantage in males. indicating a greater left hemisphere bias. lbwever. other stuiies have found a greater right ear advantage in females (Carter-Saltzaan. 1979: McKeever and VanDeventer. 1977). Simila-ly. many tachistoscopic stuiies have indicated a greater right visual field (left hemisphere) ulvantage for males in the processing of verbally presented material (e.g.. Levy and Reid. 1976: Bradshaw and Gates. 1978) and a greater left visual field (right hemisphere) advantage for males in the processing of non-verbally presented material (e.g.. Kimu'a. 1969: Rizzolatti and Rachel. 1977). Many other studies. however. did not find these sex differences (e.g.. Hanney and Boyer. 1978: McKeever and VanDeventer. 1977). Furthermore. the interwetations of experiments using tachistoscopic (Sergent. 1982) or dichotic listening (Teng. 1%1) techniques have been called into question (see also McGlone. 1980). nectroencephalegraphic (EEG) records have also been used to assess functional lateralization. bu: there are methodological woblems in the interpretations of EEG cognitive asymmetries (Donchin ~et al.. 1977: Gevins et al.. 1979). It is therefore not sunrising that some investigators find greater asymmetry in task related EEG for men (Ray et al.. 1976: Tucker. 1976). others find a greater asymmetry for females (Davidson et al....1976: Rebert and hhoney. 1978). and still others find no sex difference (Galin et al.. 1%2). Anatomical asymmetries have been described in frontal (Falzi et al.. 1982: LeMay. 1977: Heinberger et al.. 1982). tuporal (Geschwind and Devitsky. .1968: Nada et al.. 1975). occipital (Ne‘inberger et al.. 1982: Lenny. '1977). and parietal (LeMay and mlebras. 1977: Delacoste-Utnsing and Holloway. 1982b) regions. fiends toward more anatomically asymmetrical brains in males have been noted (e.g.. Debby. 1977). In suamm'y. the hypothesis that female brains function more symmetrically than male brains is countered by opinions that the reports of sex differences in functional lateralization are. in general. unreliable (Annett. 1980). unconvincing (Fairweather. 1976: Fairweather. 1980: Kinsbou'ne. 1980). or inconclusive (Ibrahall. 1973: Dryden. 1978: Martin. 1980: Sherman. 1980). mummass DIFFERENCES Simila' types of data have been used in support of the hypothesis . that left-handers as a groun show less functional asymmetry than right-headers (see Herron. 1980 for reviews). In this case as well. the neu'ophysiolegical conclusions are rarely based on neu'ophysiological measu'es: however. handedness differences are generally reported in the hypothesized direction. The study of cerebral blood flow may also address other questions of handedness differences. Harris (1980) has described early theories on the origin of handedness. The results of the resent study can address one such theory that relates hand preference to cerebral blood supply. According to this theory a greater blood supply to one hemisphere causes contralateral handedness. Althowgh this theory fell from popularity because its anatomical :remises proved faulty. cu'rent techniques to directly blood supply to the brain allow a reassessment. DIRECT MEASURES Techniques to measu'e cerebral blood flow and metabolia make it possible to assess sex and handedness differences in functional asymmetries of the brain on a direct physiological basis. Since these techniques simultaneously measu'e activity in a nuaber of relatively local regions. they provide the opportunity to localize functional asynmetries with an anatomical resolution much better than that of a cerebral hemisphere or cortical lobe. an impossibility with the indirect measu'es. Furthermore. these techniques permit analysis. beyond the scope of the resent study. of interactions among many (potentially all) brain regions. and thus will eventually fulfill their wemise to advance neu'opsycholegy beyond the simple conceptions of functional localization and hemispheric asymmetry to which the field has been technologically limited. In the early stages of their use. however. these techniques must be put to the questions of asymmetry and localization. to test and refine the ideas generated them the indirect measures. H'ohovnik et al. (1%0) exannined regional cerebral blood flow (rm?) du'ing rest in a group of young. right-handed males. They calculated interhemispheric correlations betwmen 16 pairs of homotopic regions and found positive correlations between all but one pair. These correlations were higher over pnima'y sensory regions than over secondary or tertiary (association) regions. They also chose from their saeple of 22 swbjects the five subjects scoring highest and lowest with respect to right-handedness on a handedness questionnaire. They found that mean hemispheric flow was higher to the right hemisphere for the high right-handedness group relative to the low right-handedness group. This group difference was significant at three superior precentral regions: in two of the regions there was an apparent left asymmetry for the low right-handedness group. whereas in the other region there was an apparent right asymmetry for the high right-handedness group. The results snggest that rCBF asymmetries in superior p‘ecentral regions interact with the handedness. however. tests of significance for within group asymmetries were not reported. 10 Gur et al. (1982) studied rCBF during rest and cognitive activity in a group of 62 healthy. young adults composed of right-handed males. right-handed females. left-handed males. and left-handed females. They found that solving verbal analogies or line orientation problems resulted in increased blood flow in all groups relative to rest. and that overall flow was greater during the spatial task than during the verbal task. They reported a task by hemisphere interaction reflecting a greater left hemisphere flow during the verbal task and a greater right hemisphere flow during the spatial task. They also reported a sex by handedness by task by hemisphere interaction reflecting a greater asymmetrical change from task to task for females than for males and for right-handers than for left-handers. The question of sex and handedness differences in interhemispheric ‘asymmetries or. correlations at homotopical regions within each hemisphere was not directly addressed in the paper of Our et al. (1982). The purpose of this study is to do so. INTERHEMISPHERIC CORRELATIONS AND ASTMMETRIES Correlations between homotopical regions are believed to be indicative of interhemispheric communication. Two exceptions are the situations in which homotopical regions are functionally independent yet react to a stimulus condition in similar (positive correlation) or opposite (negative correlation) ways. Subcortical structures are poor candidates to be the direct substrate of correlations since structures known to have widespread bilateral cortical efferents. specifically the locus coeruleus and the raphe nuclei. are not known to affect the cortex with any regional specificity. The claustrum has been reported 11 to have bilateral cortical efferents (Macchi et al.. 1981). but the functional significance of these connections has not been established. 0n the other hand. the role of the corpus callosum in interhemispheric communication is well established (Steele-Russell et al.. 1979). and there are abundant interhemispheric connections (e.g.. Enner and Myers. 1965). Given the anatomical facts. the possibility of the hemispheres being correlated in activity but functionally independent seems relatively unlikely. Therefore. a positive correlation. in the absence of an absolute asymmetry in flow. is to be interpreted as evidence of symmetrical (excitatory) communication between the' hemispheres. and more positive correlations for one group as evidence of more symmetrical interplay between the hemispheres for that group. lateral asymmetry is a general term signifying that one hemisphere (or regions thereof) is different from the other. There are three. not necessarily exclusive. types of functional asymmetries. thmispheric superiority is the condition in which one hemisphere participates in a function to a greater extent than the other. lbmispheric specialization is the condition in which one hemisphere has features that the other does not. lbmispheric dominance is the condition in which one hemisphere has a competitive advantage in contributing to a function that could have involved either hemisphere: this would include inhibition of a potentially interfering mnction. An asymmetry occurring with a positive correlation is consistent with superiority. an asymmetry occurring with a negative correlation is consistent with dominance. and an asymmetry that occurs with no correlation is consistent with specialization. Each of these possibilities could result in a right-left asymmetry in rCBF. 12 The laterality index and interhemispheric correlation to be calculated reflect different aspects of interhemispheric functional asymmetries. The laterality index is a measure of the percent of interhemispheric difference in the magnitude of rCBF. The interhemispheric correlation is not predictable '2 [pgiggi from the laterality index. i.e.. the two measures are conceptually independent. A positive correlation is interpreted as excitatory communication‘ between the hemispheres (symmetrical interhemispheric interplay) and a negative correlation is interpreted as inhibitory interhemispheric communication (asymmetrical interhemispheric interplay). Thus. whereas the laterality index is a measure of asymmetry in the magnitude of activity. the correlation is a measure of asymmetry in the direction of activation . M ETHOD SUBJECTS The sample of subjects was the sample of (hr et al. (1982). (he right—handed female was excluded since data free all 8 homotopical pairs of regions were not available on her. The sample used in this study was comprised of 61 volunteers ranging in age from 18 to 26. Subjects were recruited from an advertisement in the university of Rennsylvania student newspaper. Most of the subjects were undergraduate students. but that status was not one of the selection criteria. Informed consent was obtained from the subjects. and participating females signed a declaration of non-pregnancy. The subjects received $15 for participation. As a monetary incentive the subjects were informed that they would receive 25 cents for each problem they correctly solved. Payment was disbnrsed to the subjects after their testing sessions were complete. Snbjects were grouped according to sex and handedness. The groups consisted of right-handed males (n=15). right—handed females (nznu). left-handed males (n=15). and left-handed females (n:17). Sex was detenmined by inspection. The criterion for handedness classification was. the hand preferred for writing. The following personal data were recorded on each subject: age. handwriting posture. eye dominance. familial sinistrality. history of handedness change. and self-reported behavioral dominance for pointing. drawing. hammering. throwing a ball. dealing cards. using a screwdriver. and cutting with scissors. Although subtypes of left-handedness (reversed laterality. mixed laterality. pathological etiology) have been suggested. and 13 l4 left-handedness has been considered to be a continuum of non-dextrality (cf. Herron. 1980). this study will consider the group of left-handers in general. In initial step in understanding the interhemispheric relationships of cerebral blood flow that are characteristic of left-handers is to examine the matter in an undifferentiated group of left-handed subjects. Exmnnination of regional cerebral blood flow in subtypes of left-handers will be the subject of future work. PROCEDURE All rocedures and protocols have been approved by the Committee on Studies Involving Human beings and the Radioactive [rug Research (bmmittee of the university of Pennsylvania and the University Committee on Research Involving Human Subjects of Michigan State university. Regional cerebral blood flow was measured th-ee separate times on each subject. once during each of three cognitive conditions: resting -- the subject was instructed to lie still. keep eyes open. and stay awala: verbal -- the subject solved verbal analogies taken from Miller's Malogies Test (Turner. 1973): spatial -- the subjects solved Benton's Line Orientation Test (hnton et al. 1975) adapted for slide reaentation. Each cognitive condition lasted 15 minutes. and a break of .15-20 minutes separated the conditions. The order of the conditions was va'ied randally across the subjects. and a significant effect of task order was not found. The cognitive roblsas were rejected from slides onto a screen located wove the supine subjects. who indicated their anawars using a bimanually controlled flashlight arrow. The subjects' responses were recorded by the experimernter. who controlled the slide rojector from the subject's left. The verbal and spatial 15 tasks were chosen to exploit the functional biases of each hemisphere. (hr and Reivich (1980) showed that the analogies used do preferentially activate the left hemisphere in right-handed males. The Line O’ientation Test was chosen on Benton's recommendation (personal ' communication) that it is the most sensitive test for detecting right hemisphere lesions. The validity of these choices of tasks was further substantiated by the analysis of Our et al. (1982). which was repeated on the 61 subjects of this analysis: an AIDVA of rCBF data indicates a significant task by hemisphere interaction [F(2.11l1)=13.l15.p < .001] that reflects a greater left than right hemisphere increase in flow for the verbal task and a greater right than left hemisphere increase in flow for the spatial task. THE 133-XENON INHALATION TECHNIQJE The rates of regional cerebral blood flow (rCBP') were measured in all. subjects. The ideas behind this measursunent are straightforward and a brief exposition of them follows. The technical details can be found in wrist et al. (1975) and (brist and Wilkinson (1%0). Cerebral blood flow is highly coupled to glucose metabolic rate in normal brains and is therefore postulated to be an index of neuronal activity (Reivich. 19711: Raichle et al.. 1976). Regional variations in .. the distribution of blood in the brain are controlled rima'ily by the local chemical environment that reflects the state of metabolic need. Changes in the concentrations of the roducts and reactants of metabolic reactions cause local changes in the diameter of cerebral blood vessels. For example. glucose metabolise and oxidative phosphorylation result in an increased tissue concentration of ca'bon dioxide (C02) and a decreased tissue concentration of oxygen (02). 16 Vasodilation of cerebral arterioles is caused by decreased 02 or incresed C02 concentration. As a result. more blood and therefore more nutrients are brought to the area in need. In the experimental rocedure the subjects lie comfortably on their back with a mask fitted over their mouth and rnose. A mixture of room air and 133-xenon (approximate dose is SmCi/liter of air) is inhaled continuously for aproximately one minute. followed by 111 minutes breathing of normal room air. During these fifteen minutes the clearance of the radioactive xenon tYom the brain is recorded by 16 extracranial sodium-iodide crystal scintillation detectors 15 mm in surface diameter with lateral collimation 1.9 cm in diameter and 2.1 cm in length. The 133-xenon diffuses into the blood from the lungs and the tissue from the blood. but. being inert. it does not react with any of the tissue constituents. It leaves the tissue and is exhaled. The rate of rCBF is calculated from the rateof clearance of the isotope from the tissue under a particular detector. The number of radioactive counts registered at each detector is recorded in six-second intervals and may be plotted versus time (Figure 1). Rom these data a clearance curve for the isotope is generated by a least squares curve-fitting algorithm. This curve' fitting does not male use of data points that are collected during aproximately the first minute and one-half of the study to avoid the roblem of air _ passage artifact. Air passage artifact is the scattered radiation measured by extracerebral detectors due to the resence of 133-xenon in the mouth. throat. and nasopharynx. This roblanu disappears shortly after the isotope inhalation ceases. Onrve fitting begins when the count rate in the exhaled air has decreased to 20% of its maximum. 17 Three tissue types (called compartments) contribute to the clearance curve: gray matter. white matter. and extracerebral tissue (see Figure 8 of wrist et al.. 1975). Gray matter flow can be readily separated hem white matter flow. since its clearance rate is seven times that of white matter. In a two campartmental model that is commonly used. the gray matter compartment is mathematically separated from a compartment containing white matter and extracerebral tissue. The extracerebral compartment is relatively small with a relatively slow clearance rate and constitutes a minor portion of the slow perfusing compartment. The data were collected on line by a PDP-11 computer. which is used for the clearance curve analysis. One commonly used measure of rCBP is f1. or fg. the rate of rCBF in the gray matter (first) compartment in terms of ml/100 gm tissue/minute. In pathological brains or brains with a low overall rate of flow. f1 is an unreliable measure because of the roblsm of slippage. Slippage is a violation of the basic assumption of the two campartmental model: the two compartments are not as greatly different and therefore contaminate one another. To get around the roblem of slippage. flow parameters are also calculated fl'om the entire clearance curve. using a so-called menu-compartmental model. Christ's Initial Slope (IS) is a non-compartmental parameter that rovides an index of gray matter flow but is virtually unaffected by slippage. It is defined as the tangent to the clearance curve at time zero of curve-fitting for an equivalent bolus injection of 133-xenon (Obrist and Hilkinson.1980). It is referable to Risberg's Initial Slope (Risberg et al.. 1975). since the latter. using a later portion 18 of the clearance curve (minutes 2 to 3). is less purely a measure of gray matter. IS may be used for normal as well as pathological conditions. Since there are no recise criteria for defining a neurologically normal human being. normal is a default category including anyone with no known history of neurological or psychiatric roblems. It 'is therefore possible that the normal subjects in this study include some people with undiagnosed abnormality. To minimiu the potential effects of this possibility. i.e.. to use a stable measure. (brist's B parameter was used. The detectors were positioned over eight pair of homotopical regions. These detectors were attached to a helmet worn by the subjects and were oriented at angles normal to the curvature of the skull. The aproximate detector positions. as illustrated in Figure 2. were determined by a neuroradiologist using an X-ray of one subject wearing the helmet. The locations of the detectors are 1: precentral. 2: inferior recentral. 3: . middle recentral. 11: superior recentral. 5: superior posterior temporal. 6: inferior parietal. 7: superior postcentral. 8: posterior parietal-superior occipital. Due to individual variations in the size and shape of skulls and brains. the localization of the tissue seen by a detector is not recise. but it does yield a good approximation. probanly within a few millimeters. A further imrecision in localization stems them the fact that the region of tissue potentially surveyed by a detector is conical with its apex near the detector. This is not considered to be a major robleum. since most of the activity registered at adetector comes hem a roughly cylindrical volume extending through neocortex and underlying white matter and. to a diminishing extent. deeper structures. l9 STATISTICAL ANALYSIS Product-mement correlation coefficients were calculated for rCBF between homotopical regions for all groups for each of the three tasks. Therefore. for each grouping of subjects there were 211 correlation coefficients generated: one for each of. the eight homotopical pairs and eight for each of the three tasks. The subject groupings were male (0:30). female (m31). right-handers (n=Z9). left-handers (n=32). right-handed males (n215). right-handed females (mun). left-handed males (n=15). and left-handed females (n=17). Group differences between correlations were tested by ceumparing the Fisher z-transformations of the correlation coefficients. The formulas used were (z1-zZ)/[(1/n1-3)+(1/u12-3)1 for the unpartialled correlations and (z1-zZ)/[(1/n1-11)+(1/12-A)1 for the partialled correlations. It was necessary to use partial correlations because intersubject differences between detector values are much greater than intrasubject differences. In other words. total brain blood flow is a greater determinant of the absolute values of rCBF than are region to region variations in blood flow. Thus. correlations between a pair of regions will not only reflect the correlation between those particula' regions but also their mutual correlation with total brain flow. It is therefore important to statistically control «for the effect of total brain flow on the interregional correlations. Total brain flow is the mean IS at the 16 detector locations. Partial correlations between homotopical pairs were calculated in which the effect of whole brain flow was partialled out. All statistical analyses were performed using rograms of BMDP (UCLA Dept. of Riomathematics. 1979) at the Michigan State University (bmputer Laboratory. Correlations were calculated using rogram 20 BiDPéR. An analysis of va'iance with repeated measures (using rogram BiDPZV) has been performed on the total sample of 61 subjects for all th'ee tasks. using sex and handedness as between group variables. and task. hemisphere. and region the within group (repeated measures) vaiables (see Guret al.. 1982 for the analysis of the original 62 subjects). Significant interactions were found that justify the search for regional asymmetries and group differences in asymmetries. These interactions include hemisphere by region [F(7.399)=u.211. p < .001 J. hemisphere by sex [F(1.57)=11.21. p < .05]. hemisphere by hand reference [F(1,57):5.111. p < .03]. and the four way interaction of task by hemisphere by sex by hand reference [F(2. 11111)=5.96. p < .0011 1. Within group asymmetries in regional flow were assessed by paired t-tests between homotopical regions. (roup differences in regional asymmetry were assessed using a t-test ccmnparison of the laterality index [(right-left)/(right+left)] X 100. Analyses. were done using rograa B‘IDPi). RESULTS SEX DIFFEREMES Resting: correlations. The right-left correlations for the 30 . male and 31 female subjects are illustrated in Figure 3 and listed in Table 1. All the correlation coefficients were very high. between .80 and .95. well beyond the .001 level of significance. In all 8 regions the correlation coefficients were higher for females than for males. As noted above. these correlation coefficients not only reflect the interhemispheric covariance between homotopical regions but also reflect the subject to subject variability in total brain flow due to the high correlation of regional to total brain blood flow. To focus on purely regional correlations. the effects of total brain flow were controlled by computing the partial correlations of rCBF betwaen homotopical regions. partialling out total brain blood'flow. All 8 of the partial correlation coefficients for females were positive (chi square=8.0. df=1. p < .01: Figure II. Table 3). These correlations reached statistical significance at the middle recentral and superior postcentral regions. A significant negative correlation was. found for males at the middle recentral region. The sexes differed significantly at the middle recentral region: the partial correlation coefficient was more positive for the females than for the males; The incidence of positive correlations was greater for females than for males (chi square=7.3. df=1. p < .01). Among right-handed subjects. all 8 of the partial correlation coefficients for females were positive (chi square=8.0. df=1. p < .01: 21 22 Figure 5. Table 3). These correlations reached statistical significance at the inferior precentral. middle [recentral. superior posterior temporal. and superior postcentral regions. A significant positive correlation was found for the males at the superior precentral region. The sexes differed at the inferior and middle {recentral regions. where the partial correlation coefficients were more positive for females than for males. and at the superior precentral region. where the correlation was more positive for males. The incidence of positive correlations was greater for females than for males (chi square=5.3. df=1._p < .05). Imong left-handers. the partial correlation coefficients for females were positive at 7 of the 8 pairs of regions (chi squaresu.5. dfs1. p < .05: Figure 5. Table 3). tbne of these correlations reached statistical significance. There was a significant negative correlation for males at the middle precentral region. The sexes differed significantly at this region. Thus. the same sex difference in the correlation between the right and left middle urecentral regions was found for left-handers as well as right—handers. The incidence of positive correlations was again greater for females than for males (chi square:'1.3. dfz1. p < .05). as was the case for the ri'ght-handers. ., Resting: asyunetries. The group of 30 males had significantly asymmetrical rmF to the left hemisphere. i.e.. greater left than right hemisphere rCBF. at the inferior recentral and superior [recentral regions (Figure 6. Table 5). The 31 females had a significant mean Lightesymmetry at the superior [recentral region. Significant sex differences in asymmetry were found at both inferior and superior precentral regions. At the superior ure‘central region. however. the 23 sexes each had an asymmetry that was significant but in opposite directions. The manitude (absolute value) of the asymmetry was greater for males than for females at 7 of the 8 loci (chi squares-11.5. dfz1. p < .05) ‘ Among right-handers. males had significantly asymmetrical rCBF to the left hemisphere at the inferior precentral region. whereas females had no significant asymmetries (Table 6). A significant sex difference in asymmetry was found at the superior recentral region. Among left-handers. males had a significant asymmetry to the left hemisphere at the superior precentral region (Table 6). Females had no significant asymmetries. A significant sex difference in asymmetry was found at the superior precentral region. as was the case for right-handers. For both handedness groups males had greater left hemisphere rCBF at this region and females had greater right hemisphere flow. lbwever. the magnitude of the asymmetry was greater for females among right-handers. and greater for males among left-handers. although these manituie differences were not statistically significant. The mq'nitude of the asymmetry was greater for males than for females at 7 of the 8 loci (chi squarezu.5. dfz1. p < .05). but the direction of the asymmetry differed between the sexes at 6 of the 8. ._ 13523; _t_g:_s_1_c_:_ ‘ correlations. A significant positive correlation was found for females at the prefrontal and superior postcentral region (Figure 7. Table 3). A significant negative correlation occured for females at the middle [recentral region. whereas during rest they had a positive correlation at this region. The correlations for males were negative at S of the 8 loci. 1b correlations were significant for males. The sexes differed significantly at two regions: the partial 24 correlation coefficient was more positive for the females than for the males at the urethontal region. bu: it was more negative at the middle precentral region. The sex difference at the middle precentral region was opposite to that found during rest. ‘ Imong right-handed subjects. the partial correlation coefficients for females reached statistical. significance at superior preirontal and superior postcentral regions. At the middle precentral region the correlation for females was significantly negative. The correlations for males reached significance at no region. The sexes differed significantly at the weirontal and superior postcentral regions. the partial correlation coefficients were more positive for the females than for the males. Along left-handers. none of the partial correlation coefficients for males or females reached statistical significance. The sexes differed significantly at the - middle recentral region: the correlation coefficient was more negative for females. The. sex difference at this same region was opposite to that found along left-handers during rest. 15522; JEPJSE. asmnetries. The group of 30 males had significantly asymmetrical rCBF to the left hemisphere at 3 regions: inferior precentral. inferior parietal. and posterior parietal-superior occipital (Figure 8. Table 5). The 31 females had a significant mean left asymmetry. i.e.. greater left than right rCBF. at the superior postcentral region. Significant sex differences in asymmetry were found at both the superior postcentral and posterior parietal-superior occipital . 25 Among right-handers. males had significantly asymmetrical rCBF to the left hemisphere at 11 regions: inferior precentral. superior precentral. inferior parietal. and posterior parietal-superior occipital (Table 6). Females had significant left asymmetries at the superior postcentral and posterior parietal-superior occipital regions. Although the munitudes of the left asymmetries were even greater for flmnales than males at the inferior frontal and ‘inferior parietal regions. these did not reach significance due the higher standa‘d errors. A significant sex difference in asymmetry was found at the superior postcentral region. where males had a non-significant right asymmetry and females had a significant left asymmetry. Among left-handers. males had a significant asymmetry to the left hemisphere at the inferior precentral and posterior parietal-superior occipital regions. Females had no significant asymmetries. A significant sex difference in asymmetry was found at the posterior parietal-superior occipital region. Spatial £225: correlations. The partial correlation coefficients for flmnales reached statistical significance at the posterior parietal-superior occipital region. A negative correlation for males resend significance at the superior {recentral region. The sexes did not differ significantly at any region. Among right-handed subjects. the females had a significant positive correlation at the superior posterior temporal region. whereas males had a significant negative«correlation at superior postcentral region (Table 3). fine sexes did not differ significantly at any region. At 6 of the 8 regions. however. the female correlations were more positive. 26 Among left-handers. none of the partial correlation coefficients for females were significant (Table 3). At the superior precentral region left-handed males had a significant negative correlation. The sexes differed significantly at this region. Spatial £9315: asygnetries. The group of 30 males had one significant asymmetry in rCBF (Figure 8. Table 5). m1: asymmetry was at the posterior parietal-superior occipital region. and despite the fact that the cognitive task was spatial. the asymmetry was to the left hanisphere. At 6 of the other 8 regions the asymmetry was to the right hemisphere but not significant. The 31 females had no signicant asymmetry. but. the mean asymmetry at the posterior parietal-superior occipital region was also to the left. There were no significant sex differences. and. with the exception of the inferior parietal region. the asymmetries for both sexes were nearly identical. Among right-henders. males had significant asymmetrical rCBF to the left hemisphere at the posterior parietal-superior occiptal region (Table 6). Right-handed females also had a significant left asymmetry at this region and had a right asymmetry at the superior posterior temporal region. 1b significant sex difference was found. Imong left-handers. males had a significant asymmetry to the' right hemisphere at . the middle urecentral region. Left-handed females had no significant asymmetries. No significant sex difference was found for left-handers. as in the case for right-handers. iaft-handers of both sexes also showed a mean left asymmetry at the posterior parietal-superior occipital but neither were significant. Sue-nary _¢_n_i: 335 differences. Sex differences in interhemispheric partial correlations were most evident during the resting condition. 27 Females had a greater number of positive correlations. and correlations that were more strongly positive. The differences were most evident precentrally and along right-handers. Females had a more positive correlation at the middle {recentral region than males in both the right- and left-handedness groups. The pattern was simila' during the cognitive tasks. but regional sex differences were less frequent. Sex differences along left-handers at the middle recentral region were reflected in more positive correlations for females during rest but more positive correlations for males during the verbal condition. Males had -a greater number of significant regional asymmetries than females. This was especially so during the resting and verbal conditions and along the right-handed subjects. A sex difference was found at the superior precentral region: this difference was present for both handedness groups with males showing greater relative left-hanisphere flow. Taft-handed females had no significant regional asymmetry during any of the three conditions. (“NINE-188 DIFFEREKI Resting: correlations. The right-left correlations for the 29 right-handed and 32 left-handed subjects are listed' in Table 2. For the..reason mentioned above only partial correlations will be exawnined in detail. In 7 of the 8 regions the partial correlation coefficients for right-handers were positive (chi squareeu.5. (“‘31. p < .05: Figure 9.‘ Table '1). These correlations reached statistical significance at the superior posterior temporal and superior postcentral regions. The positive correlation for left-handers at posterior parietal-superior 28 occipital region is significant. The handedness groups *differed significantly at 11 regions: the inferior and middle [recentral. superior posterior temporal. and posterior parietal-superior occipital. at which the correlations for right-handers were more positive than those for left-handers. Imong female subjects. all 8 of the partial correlation coefficients for right-headers were positive (chi square:8.0. df=1. p < .01: Figure 10. Table I). These correlations reached statistical significance at the inferior precentral. middle precentral. and superior postcentral regions. The correlations for left-handers were positive at 6 of the 8 loci. with none reaching significance. The handedness groups differed significantly at the inferior and middle precentral regions. the partial correlation coefficients were more positive for the right-handers than for the left-handers. Imongmales. a significant positive correlation occurred at the superior [recentral region. None of the correlations for left-handers were significant. The handedness groups differed significantly at _the superior :recentral region. Resting: asymmetries. The group of' 32 left-handers had no significant asymmetrical rCBF (Figure 11. Table 7). The 29 right-handers had ‘a significant mean left asymmetry at the inferior precentral region. No significant handedness differences in asymmetry were found. Imong females. there were no significant asymmetries for right-handers or left-handers and no significant handedness differences (Table 8). Mean asymmetries were to the right in 7 of 8 regions for left-handers and to the left in S of 8 for right-handers (chi 29 square:fl.3. df=1. p < .05). Imong males. left-handers had a significant asymmetry to the left hemisphere at the superior recentral region (Table 8). Right-handers had a significant asymmetry at the inferior precentral region. A significant handedness difference in asymmetry was found at the superior :recentral region. the region at. which significant sex differences were found for both handedness groups (see above). 'lggbgl.§gg§: correlations. The partial correlation coefficients for right-handers were positive at 6 pairs of regions (Figure 9. Table 11). The positive correlations reached statistical significance at region the yrefiontal. and superior postcentral regions. A significant negative’correlation occurred at the middle precentral region. No correlations were significant for left-handers. The handedness groups differed significantly at the wefl'ontal and superior postcentral regions: the partial correlation coefficient was more positive for the right-handers than for the left-handers at both regions. Among fimmale subjects. partial correlation coefficients for right-handers reached positive significance at the superior wefrontal and superior postcentral regions. At the middle recentral region. the correlation for right-handers was significantly negative. lbne of the correlations for left-handers were significant. The handedness groups differed significantly at the urefl'ontal and superior [recentral regions the partial correlation coefficients were more positive for the right-handers than for the left-handers. Among males. none of' the partial correlation coefficients for either left-handers or right-handers reached statistical significance (Table 11). 30 MM: asymmetries. The group of 32 left-handers had significant asymmetrical rCBF to the left hemisphere at the inferior precentral region (Figure 12. Table 7). The 29 right-handers had a significant mean left asymmetry at the inferior {recentral. inferior parietal. and posterior parietal-superior occipital regions. No significant handedness differences in asymmetries were found during verbal stimulation. Among males. right-handers had significantly asymmetrical rCBF to the left hemisphere at 11 regions: inferior recentral. superior precentral. inferior parietal. and posterior parietal-superior occipital. Left-handers had significant left asymmetries at the inferior precentral and posterior parietal-superior occipital regions. Although the munituie of the left asymmetries was even greater for left-handers than right-handers at the inferior precentral and inferior parietal regions. these did not reach significance due the higher standad errors. It) significant handedness difference was found for males. Imong females. right-handers had significant asymmetry to the left hemisphere at the superior postcentral and posterior parietal-superior occipital regions. left-handers had no significant'asymmetries. A significant handedness difference in asymmetry was found at the posterior parietal-superior occipital region. The mgnitude of the asymmetry was greater for right-handers than for left-handers at 6 of the 8 loci. Spatial £295: correlations. The positive correlation coefficients for right-handers reached statistical significance at the urefl'ontal and the posterior parietal-superior occipital regions. and a 31 negative correlation was significant at the superior {recentral region (Figure 9. Table 18). None of the correlations for left-handers were significant. Tho significant handedness differences were found: at the :reirontal region. the correlation for right-handers was more positive than for left-handers. whereas at the superior postcentral region. the correlation for right-handers was more negative. At 6 of the 8 regions right-handers had a more positive correlation than left-handers. Among female subjects. the right-handers had significant positive correlations at the superior posterior temporal and posterior parietal-superior occipital regions. whereas left-handers had no significant correlation (Table 11). The handedness groups differed at the posterior superior temporal reg’ion winere right-handers had a more positive correlation. Along males.there was a significant negative correlation at the superior postcentral region for right-handcrs (Table 11). At the superior recentral region the left-handed males had a significant negative correlation. The handedness groups differed significantly at the :reihontal and superior postcentral regions: right-handers were more positive at the :ren‘ontal and left-headers mor'e positive at the superior postcentral region. Spatial pp_s_1_:_:_ asymetries. The group of 29 right-handers had one significant asymmetry in rCBF (Figure 12. Table 7). This asymmetry was at the posterior parietal-superior occipital region and despite the fact that the cognitive task was spatial the asymmetry was to the left hemisphere. At 6 of the other 8 regions the asymmetry was to the right hemisphere but not significant. The 32 left-handers had significant 32 asymmetries to the right at pren'ontal and middle recentral regions. The mean asymmetry at the posterior parietal-superior occipital region was to the left as was the case for right-handers. but it was not significant. Iaong right-handers. males had a significant asymmetry to the left hemisphere at the posterior parietal-superior occipital region (Table 8). left-handed males had a non-significant left asymmetry at the posterior parietal-superior occipital region and had a right asymmetry 'at the middle [recentral region. A significant handedness difference was found at this region. Among females. right-handers had a significant asymmetry to the right hemisphere at the superior posterior temporal region. Female left-handers had no significant asymmetries. No significant handedness difference was found for females. Females of both handedness groups show a mean left asymmetry at the posterior parietal-superior occipital region. bub this was significant only for right-handers. Sumary pg handedness differences. Handedness differences in interhemispheric partial correlations were also most evideunt during the resting condition. Right-handers had a greater number of positive correlations. and correlations that were more' positive. The differences were :resent both precentrally and postcentrally and were more evident anong right-handers. The pattern was similar during the cognitive tasks. bub handedness differences were less frequent. Right-handers had a greater number of significant regional asymmetries than left-handers. This was especially so during the resting and verbal conditions and along the male subjects. left-handed females had no significant regional asymmetry during any of the three conditions. 33 Figure. 1. ample of 133-Xe clearance curve. Agraph of the radioactive counts at an extracranial detector (thousands of counts) along the ordinate versus time (minutes) along the ascissa. The points are the result of a six-second sanpling interval. The curve-fitting was performed by a canputeriaed least-squares algoritlmn after air passage artifact had became negligible. The data in this figure were taken Rom a young male who was notasubject in this stuwdy: they are used for illustrative purposes only. 34 FIGURE 1 COUNTS 5K NARRCH PSYCH CONTROL W _ 21-OEC-82 TEST Q 5 .3. HEAD CURUE k 2 4K- -‘ 3K:- ZKH , 1K! . I I I I I T I I I I I l I j I 1 2 3 4 5 6 7 8 9 1B 11 13 14 15 . TIME IN MINUTE 35 Figure 2. Location of detectors. The approximate location of the tissue seen by the extracranial detectors. The detectors were attached to a helmet worn by the subjects and oriented at angles normal to the curvature of the skull. For ease of illustration only one hemisphere is depicted. but the locations were over mmotopical regions of each hemisphere. The locations of the detectors were 1: prefrontal. 2: inferior precentral. 3: middle :recentral. l1: superior recentral. 5: posterior superior temporal. 6: inferior parietal. 7: superior postcentral. 8: posterior parietal-superior occipital. 36 mmOHUMHwD mo Z9230... N mmDOHm 37 Figure 3. Sex differences in correlations during rest. The right-left correlations for the 30 male and 31 female subjects are illustrated. All the correlation coefficients were very high. between .80 and .95. well beyond the .001 level of significance. In all 8 regions the correlation coefficients were higher for females than for males (chi square-:8. df=1. p < .01). FIGURE 3 1.0 0.9 0.6 0.7 0.6 0.5 0.4 0.3 0.0 38 sex DIFEERENCES IN CORRELATIONS DURING REST l H MALE 0---O FEMALE i l L l l l 12 3 4 5 5 7 DETECTORS P= .001 39 Figure '1. Sex differences in partial correlations during rest. All 8 of the partial correlation coefficients for females were positive (chi square-8.0. df=1. p < .01). These correlations reached statistical significance at detectors 3. middle precentral region. and 7. superior postcentral region. The correlations for males were negative at 5 of the 8 loci. reaching significance at detector 3. the middle [recentral region. The sexes differed significantly at detector 3: the partial correlation coefficient was more positive for the females than for the males. At 5 of the resaining 7 regions. the sexes differed in the same direction. The incidence of positive correlations was greater for females than for males (chi square=7.3. df=1. p < .01). FIGURE 4 PARTIAL I‘ 40 SEX DIFFERENCES IN PARTIAL CORRELATIONS DURING REST ‘°° ‘ o--o FEMALES " - I—I MALES 3' nun SEx ourr. P<.01 ‘ '5” 52‘ “9-901 I, \ P301 .4— *.L .l \ “(3.05 /M\ . _ .’l ' _ r ~ 2n— -6 . \V/D‘so _ 0.0 ' _ r __.4__ _.P'.05 V ’ P-.01 -.6- . - -3_. _ -w.o u 1 ° 1 u n u u w 2 3 4. s 6' 7 e DETECTORS 41 Figure 5. Sex differences in partial correlations during rest: handedness groups. Among right-handed subjects. all 8 of the partial correlation coefficients for females were positive (chi square=8.0. df=1. p < .01). These correlations reached statistical significance at detectors 2. inferior precentral. 3. middle precentral. S. superior posterior temporal. and 7. superior postcentral. The correlations for males were negative at 11 of the 8 loci. and positive at 11 of the 8: reaching significance at detector 11 where a positive correlation was found. The sexes differed significantly at detectors 2 and 3. where partial correlation coefficients are more positive for the females than for the males. and at detector '1. where the correlation was more positive for males. At all of the 5 remaining regions. the female correlations were more positive. The incidence of positive correlations was greater for females than for males (chi square=5.3. df=1. p < .05). Among left-handers. the partial correlation coefficients for females were positive at 7 of the 8 pairs of regions (chi squares-”.5. df=1. p < .05). Ibne of these correlations reached statistical significance. The correlations for males were negative at S of the 8 loci. reaching significance at detector 3. the middle precentral region. The sexes differed significantly at detector 3. Thus. the sane sex difference in the correlation between. the right and left middle precentral regions was found for left-handers as well as right-handers. At 5 of the remaining 7 regions. the sexes differed in a similar manner. The incidence of positive correlations was again greater for females than for males (chi square=u.3. df=1. p < .05). as was the case for the right-handers. ' ‘- FIGURE 5 PARTIAL r 42 SEx DIFFERENCES IN PARTIAL CORRELATIONS DURING REST w. LEFT-RAIDERS 0.1 0,5 .. é 1? <. 43 Figure 6. Sex differences in asymmetry during rest. The group of 30 males had significantly asymmetrical rCBF to the left hemisphere at regions 2. inferior precentral. and II. superior precentral. The 31 females had a significant mean right asymmetry. at region 11. Significant sex differences in asymmetry were found at both detectors 2 and 11. At detector 11 the sexes each have a significant asymmetry. but in opposite directions. 'nne magnitude (absolute value) of the asymmetry was greater for males than for females at 7 of the 8 loci (chi square-=45. df=1. p < .05). The direction of the asymmetry differed between the sexes at 5 of the 8 regions. See Table 3 for sex differences anong handedness groups during the . cognitive tasks. 44 FIGURE 6 4 Sex Differences in Asymmetry During Rest f ' ' . asymmetry. aex'differenoes _ * p<0.05 W p<0.01 8. 2 - 1. ea p<0.01' m p<0.m1 — A ' * ' / \L l .n L , \ / ELI-g: 0 Q- - ~~L \\vL// \\l _ x l o '0 .. .E a 2 g . . H a male 3 P. ** Dfemale -4 — W m I I I I I I l I 2 4 6 8 Detectors 45 Figu~e 7. Sex differences in partial correlations. Resting: The results are as in Figu'e 11. Verbal: The partial correlation coefficients for females were positive at '4 pairs of regions and negative at 14 pairs. The positive correlations reached statistical significance at detector 1. the prefrontal region. A significant negative correlation occurred at detector 3. the middle precentral region. whereas during rest the females had a positive correlation at this region. The correlations for males were negative at 5 of the 8 loci: no correlations were significant for males. The sexes differed significantly at detectors 1 and 3: the partial correlation coefficient was more positive for the females than for the males at the prefrontal region. whereas it was more negative at the middle precentral region. The sex difference at the middle precentral region was opposite to that found during rest. At half of the remaining 6 regions. the females had more positive correlations than males: the Opposite was true for the other three regions. Spatial: The partial correlation coefficients for females were positive at 5 of the 8 pairs of regions. These correlations reached statistical significance at detector 8. the posterior parietal-superior occipital region. The correlations for males. however. were negative at 5 of the 8 loci. reaching significance at detector 11. the superior precentral region. The sexes did not differ significantly at any detector. At 5 of the 8 regions. however. females had a mo FIGURE 7 03 - Sex Differences in Partial Correlations Resting I Verbal ' Spatial I— Q I ' I I" I \ 04- ;I l X I I / \\ R\I p d ‘3’, d I h 0 “a I 1': m _ D. I —04 V . - sex difference | at p<0.05 .l male -03 .. “‘ p<0.01 Dfemale I I uuuuuuuuuuuuuula illllglglg 2 4 6 8 2 4 6 2 .4 6 8 Detectors p - 0.05 p- 0.01 47 Figure 8. Sex differences in asymmetries. Resting: The results are as in Figure 6. Verbal: The groUp of 30 males had significantly asymmetrical rCBF to the left hemisphere at 3 regions: 2. inferior precentral. 6. inferior parietal. and 8. posterior parietal-superior occipital. The 31 females had a significant mean left asymmetry at region 7. superior postcentral. Significant sex differences in asymmetry were found at both detectors 7 and 8. The magnitude of the asymmetry was greater for males than for females at 5 of the 8 loci. and the direction of the asymmetry differed between the sexes at 2 of the 8. Spatial: The group of 30 males had one significant asymmetry in rCBF. 'nnis asymmetry was at detector 8. posterior parietal-superior occipital. and despite the fact that the cognitive task was spatial the asymmetry was to the left hemisphere. At 6 of the other 8 detectors the asymmetry was to the right hemisphere but not significant. The 31 females had no significant asymmetry. but the mean asymmetry at detector 8 was also to the left. There were no significant sex differences. and except at detector 6 the asymmetries for both sexes were nearly identical. 48 FIGURE 8 " fleeting ! Verbal ! Spatial 82 P X :n 3‘. cs OI g 'D .5 E-z 2 3 -4- 1:11 I l ‘ . ..I w-p