Cranial Capacity and Performance on Delay-Response TaskCorrelated With Principal Sulcus Length in Monkeys

JOHN C. REDMOND, JR.*Department of Anthropology, University at Albany-StateUniversity of New York, Albany, New York 12222

KEY WORDS working memory; primate brain evolution;comparative neurology

ABSTRACT During the process of evolution, a selective advantage mayhave been gained by organisms that had the ability to utilize mentally storedinformation of a stimulus rather than the stimulus itself. The ability totemporarily store and mentally operate on stimulus information is oftentermed ‘‘working memory.’’ Within the neocortex of primates, the functionalanatomic subdivision surrounding the principal (rectus) sulcus plays animportant role in modulating the performance of delay-response tasks inmonkeys (representing working memory). However, it appears that no studyhas investigated the direct relationship between the length of the principalsulcus and performance on a delay-response task. Therefore, this paperinvestigates the relationships between principal sulcus length and perfor-mance on delay-response tasks. However, to control for the effect of overallbrain size on this relationship, cranial capacity is analyzed with bothprincipal sulcus length and delay-response performance. Results support aconsistent and significant correlation between principal sulcus length andperformance on delayed-response tasks in a variety of Old World and NewWorld monkeys. Principal sulcus length is also significantly correlated withcranial capacity; however, cranial capacity is not significantly correlated withperformance on delayed-response tasks. The results of this investigationprovide a method for analyzing cranial capacity and working-memory abili-ties in select primates based on principal sulcus length, and may prove usefulfor interpreting endocasts in the primate fossil record. Am J Phys Anthropol109:33–40. r 1999 Wiley-Liss, Inc.

Organisms that are neurologically ca-pable of temporarily storing and mentallyoperating on stimulus information of ‘‘real-world’’ stimuli in order to make future choicespossess working memory (i.e., the tempo-rary storage of information in the brain).From an evolutionary perspective, the abil-ity to make choices based on mentally storedstimuli, rather than on continual exposureto the stimuli themselves, is posited to be amajor selective advantage.

The ability to use working memory forgoal-based decisions is intimately tied toproperties of the prefrontal cortex. In pri-

mates, the prefrontal cortex has been impli-cated in the regulation of a number of com-plex behaviors including, attention, syntheticreasoning and planning, spatial orientation,global personality features, and workingmemory (Brozoski et al., 1979; Fuster andAlexander, 1971; Jacobsen, 1936). It is notsurprising, therefore, that the prefrontalcortex of the primate brain is highly differen-tiated, allowing for the regulation of a vari-

*Correspondence to: John C. Redmond, Jr., Department ofAnthropology, University at Albany-State University of NewYork, Albany, NY 12222. E-mail: redmond@empireone.net

Received 12 May 1998; accepted 16 February 1999.

AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 109:33–40 (1999)

r 1999 WILEY-LISS, INC.

ety of behaviors via multiple functional-anatomic subdivisions (Goldman-Rakic,1987). With recent advances in the neurosci-ences, a more definitive picture of the func-tional neuroanatomic properties of the pre-frontal cortex is emerging.

One functional-anatomic subdivision dem-onstrated to be important in the processingof spatial working memory and performanceon delay-response tasks in monkeys is thedorsolateral portion of the prefrontal cortex(Goldman and Rosvold, 1970; Goldman-Rakic, 1987; Sawaguchi and Goldman-Ra-kic, 1991). Historically, ablation studies ofthe 1930s, 1940s, and 1950s helped identifythis general area of the prefrontal cortex asimportant for the regulation of workingmemory (Pribram, 1950; Mishkin, 1957).More recently, Levy and Goldman-Rakic(1997) were able to isolate functional area 46(the area surrounding the principal (rectus)sulcus) as the specific cortical area involvedin regulating performance on delay-responsetasks in monkeys by selectively placing le-sions to specific areas of the dorsolateralportion of the prefrontal cortex and testing

subsequent performance (Fig. 1; modifiedafter Walker, 1940). In humans, Courtney etal. (1998) showed spatial working memoryto reside just superior and posterior to area46 in the superior frontal sulcus of thecortex. They suggest that this shift in posi-tion of spatial working memory during pri-mate brain evolution is the result of expan-sion in the anterior portions of the lateralprefrontal cortex, and may be related to theaddition of new cognitive areas in humans.Consequently, this information helps ex-plain the previously unrecognized homolo-gous functional areas between humans andmonkeys.

Delay-response tasks represent a subject’sability to hold information on-line (Sawagu-chi and Goldman-Rakic, 1991; Williams andGoldman-Rakic, 1995). This means that dur-ing a delay-response task, working memoryfunctions by mentally operating on a previ-ously given stimulus, instead of on an ongo-ing exposure to the stimulus itself, to aid indecision-making processes. Put simply, thesubject must remember the task at hand.

Fig. 1. Area 46 is situated around the principal sulcus in the monkey cerebral cortex (modified afterWalker, 1940).

34 J.C. REDMOND, JR.

Most experimental delay-response proce-dures on non-human primates have subjectsmanipulate visual-spatial information, whichthe subject is required to remember in orderto perform the task correctly. In a simpletask, the subject is presented with either thereward (direct method) or an item thatthrough conditioning comes to represent areward (indirect method). These direct andindirect rewards are hidden in one of twolocations. After a given period of time (de-lay), the subject is allowed to retrieve theitem and a correct response results in areward. The subsequent performance abilitycan then be assessed.

The Diamond-Jane strategy models behav-ioral functions across geological time byinferring behavior from comparative neuro-anatomical data in extant species that havebeen experimentally tied to those behaviors(Masterton and Skeen, 1972). This strategyseeks to correlate specific structural changesin the brain with differences in behavioralperformance. For that reason, this methodol-ogy is well-suited for formulating hypoth-eses about changes in behavior over time,and subsequently for aiding in phylogeneticinterpretations about extinct fossil pri-mates.

The principal sulcus appears in nearly allmonkeys, and its length may be comparedacross primate species (Falk, 1978, 1979).Specifically, the principal sulcus coursesthrough area 46, which is fundamentallyrelated to carrying out delay-response activi-ties (Goldman-Rakic, 1987). By applying theDiamond-Jane strategy to an analysis ofarea 46 in primates, it is reasonable tohypothesize that a correlation exists be-tween the length of the principal sulcus andperformance on delay-response tasks in mon-keys. However, because increases in thegeneral size of the brain as a whole mayeffect delay-response performance, overallbrain size must also be considered (Finlayand Darlington, 1995; Gibson, 1998). Accord-ingly, the intent of this study was to investi-gate the relationships between principal sul-cus length, performance ability on delay-response tasks, and cranial capacity in avariety of Old World and New World mon-keys.

MATERIALS AND METHODS

Latex endocasts with corresponding cra-nial capacities representing four genera ofOld World and two genera of New Worldprimates were measured to determine prin-cipal sulcus length (Table 1). Only endocaststhat reproduced good detail and were resil-ient enough to maintain their shape duringmeasurement were included in the study. Todetermine principal sulcus length, a smallpiece of dental floss was aligned along theentire length of the principal sulcus, re-moved, straightened, and measured utiliz-ing a sliding caliper.

Sample sizes ranged from 3–8 individualsper genus. Because t-tests of right and lefthemispheres for the selected genera did notindicate any significant difference, both sideswere included in the analysis. The finalmean-length measurement was recorded asthe principal sulcus size for each primategenus. Cranial capacities provided by Dean

TABLE 1. Measured principal sulcus lengthsand cranial capacities1

Primate genera

Principalsulcus

length (cm)

Cranialcapacity

(cm3)

Old World Papio (N 5 5) 2.45 1491.91 1402.26 1622.48 1972.00 163

Mandrillus (N 5 4) 2.18 1522.14 1751.55 1361.82 163

Macaca (N 5 5) 1.67 921.67 941.72 1061.84 951.83 99

Cercopithecus (N 5 8) 0.65 670.98 801.18 640.95 601.10 821.40 811.35 841.00 62

New World Lagothrix (N 5 3) 1.18 950.95 900.98 103

Cebus (N 5 5) 0.70 690.84 —0.82 681.11 701.14 100

1 Principal sulcus lengths are mean values of right and left sulciper specimen.

35PRINCIPAL SULCUS AND DELAY-RESPONSE

Falk were obtained with mustard seed atthe time the endocasts were cast (Falk,personal communication). A one-way analy-sis of variance (ANOVA) was applied toidentify significance between genera in bothprincipal sulcus length and cranial capacity.A regression analysis utilizing raw data wasperformed to determine the relationship be-tween principal sulcus length and cranialcapacity. Finally, correlation and regressionanalyses utilizing mean data were per-formed to investigate the relationship be-tween performance on delay-response taskswith principal sulcus length and cranialcapacity (Table 2).

Raw data for delay-response task perfor-mance was taken from two previously pub-lished studies (Harlow et al., 1932; Maslowand Harlow, 1932). Because solutions todelay-response tasks can utilize availablespatial as well as nonspatial cues to coverthe delay period, variations in delays be-tween different studies of the same genusmay be partly a function of methodology.However, performances can be compared incases where studies used similar methodol-ogy (Michels and Brown, 1959).

Behavioral data were compiled for thegenera listed in Table 1. Delay periods fortask performance ranged from zero to greaterthan 180 sec, denoted by the following catego-ries: 0, 5, 15, 30, 60, 120, and 180 (in sec).Within each primate genus, the percentagecorrect for each individual subject was re-corded and averaged for each delay category.Since a comparison of overall ability wasdesired, an overall score was computed foreach genus. Scores were calculated utilizingthe scale in Table 3, where the percentage of

correct responses for a given category wasconverted to the score for that category.Then the scores for each category were addedup to represent the overall score for eachgenus. For example, if a genus averaged95% correct responses for the 5-sec-delaycategory, then a score of 9 would be assigned.Subsequently, all category scores were addedto equal the total score for each genus.Statistics used to analyze behavioral dataconsisted of a multivariate analysis of vari-ance (MANOVA) for overall differences be-tween genera across all times, and indi-vidualANOVAs for between-genera differencesat individual delay times.

RESULTS

In the analysis of delay-response perfor-mance, a single MANOVA across all delaytimes and including all genera was signifi-cant (P , 0.05). To further analyze delay-response performance, individual ANOVAswere calculated for each delay time betweengenera. Figure 2 shows delay-response per-formance plotted as an average percent cor-rect across all delay times. Significance fordelay times was found between genera at 5,15, 30, and 60 sec (P , 0.05).

Differences between genera for both prin-cipal sulcus length and cranial capacity wereevaluated using individual ANOVAs. Forprincipal sulcus length differences betweengenera, a calculated F value of 88.8 wasfound significant at P , 0.01, with a criticalF of 2.3. Likewise, differences between gen-era for cranial capacity produced a signifi-cant F value of 41.6, with critical F of 2.6(P , 0.01).

TABLE 2. Computed scores and means of principalsulcus length, cranial capacity, and delay-response

task performance1

Primate generaPSL (cm)

meanCC (cm3)

meanDelay-response

score

Old WorldPapio 2.23 162 54Mandrillus 1.92 157 44Macaca 1.74 97 40Cercopithecus 1.07 73 34

New WorldLagothrix 1.03 96 23Cebus 0.92 75 16

1 PSL, principal sulcus length; CC, cranial capacity.

TABLE 3. Score conversions1

Percentcorrect Score

100–96 1095–91 990–86 885–81 780–76 675–71 570–66 465–61 360–56 255–51 1

1 The mean percent correct scored by a genus in a given delaytime is assigned the corresponding score. The scores aresubsequently added across delay times to produce an overallscore for that group.

36 J.C. REDMOND, JR.

The first regression analysis evaluatedthe relationship between principal sulcuslength and cranial capacity. Plotting cranialcapacity as a function of principal sulcuslength produced a significant R2 of 0.75 witha standard error of 20.0 (P , 0.01; Fig. 3).Next, to examine the relationship betweenprincipal sulcus length and delay-responsetask performance, a regression analysis plot-ting delay-response score as a function ofprincipal sulcus length was utilized. Theresulting calculation produced a significantR2 of 0.88, with a standard error of 5.30(P , 0.01; Fig. 4). Finally, a regression equa-tion plotting delay-response score as a func-tion of cranial capacity for each genus wascalculated to judge the effect of overall brainsize on delay-response task performance. Anonsignificant R2 of 0.63 with a standarderror of 9.5 resulted. However, because thecomputed correlation coefficient between cra-nial capacity and delay-response task scorereached near-significance (P , .06), a test was

conducted to control for the effect of overallbrain size on the relationship between prin-cipal sulcus length and delay-response score.A partial correlation between principal sul-cus length and delay-response score, control-ling for cranial capacity, produced a signifi-cant correlation coefficient of 0.993 (P , 0.01).Correlation coefficients for the reported rela-tionships are listed in Table 4.

DISCUSSION AND CONCLUSIONS

In monkeys, the ability to use workingmemory is intimately tied to the propertiesof the prefrontal cortex, and more specifi-cally, to the area around the principal sulcus(Goldman-Rakic, 1987). The finding thatprincipal sulcus length is significantly corre-lated with delay-response performance,while cranial capacity is not, may be due tothe specific role of area 46 in keeping taskson-line in higher primates. In fact, given thepresent state of knowledge about the localiza-

Fig. 2. Delay-response performance plotted as mean percent correct across delay times for all genera.*Denotes significant differences between genera for the delay times: 5, 15, 30, and 60 sec (ANOVA,P , 0.05).

37PRINCIPAL SULCUS AND DELAY-RESPONSE

tion of neurological substrates for delay-response performance relative to the princi-pal sulcus, the stronger correlation observedbetween delay-response performance andprincipal sulcus length than with overallbrain size is not particularly surprising (Bart-ley et al., 1997; Courtney et al., 1998). Becauseneurons along the principal sulcus are knownto influence spatial working-memory ability,small increases in cortical tissue in this areamay increase working-memory capacity. Con-versely, small increases in overall brain sizemay have less direct effect on the size of theprincipal sulcus and, consequently, on work-ing memory. Nevertheless, these results re-main tentative. Ideally, an expanded studyutilizing larger sample sizes representing agreater number of primate species, whereprincipal sulcus lengths and brain sizes aremeasured using, for example, noninvasivemedical imaging technology on the actualprimates performing the delay-responsetask, would provide greater support for thereported relationships.

Without knowing the detailed neuronalworkings of a variety of animal species

(including humans), gross anatomical com-parisons are compromised. However, certainbehavioral tests have been devised, which,with ever-increasing accuracy and experi-mental manipulation, are able to establishstrong correlations between functionally ac-tive brain regions and specific behaviors(Goldman-Rakic, 1987; Damasio et al., 1996;Carpenter et al., 1999). As the revolution inneuroscience continues, the level of preci-sion in the identification of the neurologicalsubdivisions responsible for specific behav-iors enables the application of these findingsto a variety of studies. Given that specificstructural areas are related to specific behav-iors (Smith and Jonides, 1999; Welker andCampos, 1963), it becomes possible to exam-ine these correlations in light of phyloge-netic relationships. Consequently, one mayspeculate that as a behavior becomes morecomplex over time, corresponding brain ex-pansion of the area controlling that behaviorshould also increase.

One preliminary application of the pre-sent results permits an estimation of cranialcapacity and ability on delay-response tasks

Fig. 3. Line fit plot of cranial capacity plotted as a function of principal sulcus length. Data pointsrepresent individual specimens.

38 J.C. REDMOND, JR.

in extinct fossil primates that exhibit simi-lar brain morphologies. For example, Doli-chopithecus arvernensis has been describedas having a modern cercopithecine brainmorphology similar to that of Papio (Radin-sky, 1974). By measuring and inserting theprincipal sulcus length of this specimen intothe regression equation for cranial capacity,a predicted cranial capacity value can becomputed. Likewise, by plugging in the mea-sured principal sulcus length into the regres-sion equation for delay-response performance,a predicted delay-response performance scorecan be derived. When computed this way,Dolichopithecus’ predicted cranial capacityis 152 cc, only 2 cc different from the 150 cc

estimated by Radinsky (1974). The pre-dicted delay-response score for Dolichopithe-cus is 50, implying a performance level at orabove the level of Papio. Thus, these regres-sion equations may prove useful for predict-ing cranial capacity and spatial working-memory abilities in other extinct primatesthat share similar brain morphology. More-over, the length of the principal sulcus mayprove useful in phylogenetic assessments ofextinct primates where only a fragmentaryfossil record is available.

The strategy of applying knowledge of theneuroanatomical bases that are known tounderlie specific behaviors in higher pri-mates will become more important in inter-preting evolutionary changes in fossils asmore knowledge of these structures comes tolight. These data should facilitate the inter-pretation of the fossil record where specificdifferences in the abilities of our ancestorsmay be hypothetically identified, and as-sessed. Although working memory operatesover seconds, its significance is delineated

Fig. 4. Line fit plot of delay-response score plotted as a function of principal sulcus length. Data plotsrepresent computed means for each genera.

TABLE 4. Correlations among principal sulcal lengths,delay-response scores, and cranial capacities1

Components Correlation coefficient P value

PSL Dly-rsp. 0.94 0.01PSL CC 0.86 0.01CC Dly-rsp. 0.80 0.061 PSL, principal sulcus length; Dly-rsp., delay-response score;CC, cranial capacity.

39PRINCIPAL SULCUS AND DELAY-RESPONSE

over millennia. The differential abilities ofprimate genera to hold information on-linemay help shed light on the evolutionaryhistory of behavioral traits such as reason-ing, and language comprehension (Wickel-gren, 1997), and would therefore prove use-ful in the interpretation of proposed modelsof primate evolution.

ACKNOWLEDGMENTS

The author thanks Dr. Dean Falk, who isthe recipient of NSF grant SBR-9729796, forher tireless reviews of this manuscript andfor access to her endocast collection. Thanksalso go to Dr. Tim Gage, Audry Choh, Bar-bara Melinger, Dan White, Martin Solano,Shawn Phillips, and Lee Massey for theirexcellent reviews and comments on themanuscript. Most of all, I thank my parents,wife, and daughter Hannah, for their sup-port which has made this work possible.

LITERATURE CITED

Bartley AJ, Jones DW, Weinberger DR. 1997. Geneticvariability of human brain size and cortical gyralpatterns. Brain 120:257–269.

Brozoski TJ, Brown RM, Rosvold HE, Goldman PS.1979. Cognitive deficit caused by regional depletion ofdopamine in prefrontal cortex of rhesus monkey.Science 205:929–931.

Carpenter AF, Georgopoulas AP, Pellizzer G. 1999. Mo-tor cortical encoding of serial order in a context-recalltask. Science 283:1752–1757.

Courtney SM, Petit L, Maisog JM, Ungerleider LG,Haxby JV. 1998. An area specialized for spatial work-ing memory in human frontal cortex. Science 279:1347–1351.

Damasio H, Grabowski TJ, Tranel D, Hichwa R, Dama-sio AR. 1996. A neural basis for lexical retrieval.Nature 380:499–505.

Falk D. 1978. Brain evolution in old world monkeys. AmJ Phys Anthropol 48:315–320.

Falk D. 1979. Comparative study of the endocranialcasts of New and Old World monkeys. In: Chairelli B,Ciochon RL, editors. Evolutionary biology of the NewWorld monkeys and continental drift. New York:Plenum.

Finaly BL, Darlington RB. 1995. Linked regularities inthe development of evolution of mammalian brains.Science 268:1578–1584.

Fuster JM, Alexander GE. 1971. Neuron activity relatedto short-term memory. Science 173:652–654.

Gibson KR. 1998. Bigger is better: primate brain size inrelationship to cognition. Am J Phys Anthropol [Suppl]26:101–102.

Goldman P, Rosvold HE. 1970. Localization of functionwithin the dorsolateral prefrontal cortex of the rhesusmonkey. Ep Neurol 27:291–304.

Goldman-Rakic PS. 1987. Circuitry of primate prefron-tal cortex and regulation of behavior by representa-tional memory. In: Mountcastle VB, Plum F, GeigerSR, editors. Handbook of physiology, volume V: thenervous system. Bethesda, MD: American Physiologi-cal Society. p 373–418.

Harlow HF, Uehling H, Maslow AH. 1932. Comparativebehavior of primates: I. Delayed reaction tests onprimates from the lemur to the orangutan. J CompPsychol 13:313–343.

Jacobsen CF. 1936. The functions of the frontal associa-tion areas in monkeys. Comp Psychol Monogr 13:3–57.

Levy R, Goldman-Rakic PS. 1997. Segregation of work-ing memory functions in prefrontal cortex in pri-mates. Soc Neurosci Abstr 23:194.19.

Maslow AH, Harlow HF. 1932. Comparative behavior ofprimates: II. Delayed reaction tests on primates atBronx Park Zoo. J Comp Psychol 81:53–61.

Masterton B, Skeen LC. 1972. Origins of anthropoidintelligence: prefrontal systems and delayed alter-ation in hedgehog, tree shrew, and bush baby. J CompPhysiol Psychol 81:423–433.

Michels KM, Brown DR. 1959. The delayed-responseperformance of raccoons. J Comp Physiol Psychol52:737–740.

Mishkin M. 1957. Effects of small frontal lesions ondelayed alteration in monkeys. J Neurophysiol 20:615–622.

Pribram KH. 1950. Some physical and pharmacologicalfactors affecting delayed response performance ofbaboons following frontal lobotomy. J Neurophysiol13:373–382.

Radinsky L. 1974. The fossil evidence of anthropoidbrain evolution. Am J Phys Anthropol 41:15–28.

Sawaguchi T, Goldman-Rakic PS. 1991. D1 dopaminereceptors in prefrontal cortex: involvement in workingmemory. Science 251:947–949.

Smith EE, Jonides J. 1999. Storage and executiveprocesses in the frontal lobe. Science 283:1657–1661.

Walker AE. 1940. A cytoarchitectural study of the pre-frontal area of the macaque monkey. J Comp Neurol73:59–86.

Welker WI, Campos GB. 1963. Physiological significanceof sulci in somatic sensory cerebral cortex in mam-mals of the family Procyonidae. J Comp Neurol 120:19–36.

Wickelgren I. 1997. Working memory linked to intelli-gence. Science 275:1581.

Williams GV, Goldman-Rakic PS. 1995. Modulation ofmemory fields by dopamine D1 receptors in prefrontalcortex. Nature 376:572–575.

40 J.C. REDMOND, JR.