HORMONES

AND

BEHAVIOR

25,

367-381 (1991)

The Association between Testosterone, Sexual Arousal, and Selective Attention for Erotic Stimuli in Men GERIANNE

M. ALEXANDER*”

AND BARBARA

B. SHmwm*9

*Department of Psychology, McGill University, Montreal, Canada, and TDepartment of Obstetrics and Gynecology, SMBD-Jewish General Hospital, Montreal, Canada Twenty-six, eugonadal men between the ages of 18 and 27 participated in this investigation of the relationship between sexual arousal, testosterone (T) levels, and the processing of sexual information. At each of the two test sessions,subjects gave a blood sample, listened to an erotic or neutral priming audiotape, and completed a dichotic listening task designed to assessselective attention for sexual stimuli. Subjective levels of sexual arousal to the audiotape and sexual attitudes and sexual experience were assessed by self-report measures. Contrary to our hypothesis, there was no relationship between levels of free T and the strength of the selective attention bias for sexual stimuli. However, men who were more distracted by the sexual material in the task reported higher levels of sexual arousal to erotic imagery than men who were less distracted by the sexual material in the task (P < 0.01). Moreover, men who were more sexually aroused by the erotic audiotape made significantly less shadowing errors in the erotic prime condition then they did during the neutral prime condition (P < 0.05). There was a negative association between T and shadowing errors in the erotic prime condition (P < 0.05). These results suggest that lower thresholds for sexual arousal are associated with a greater bias to attend to sexual information and that T may have effects on cognitive-motivational aspects of sexual behavior by enhancing attention to relevant stimuli. 0 1991 Academic Press. IX.

Studies of testosterone (T) replacement therapy for the treatment of hypogonadal men have demonstrated that T enhances overall sexual activity (Davidson, Camargo, and Smith, 1979), the frequency of sexual thoughts, and the level of subjective arousal (O’Carroll, Shapiro, and Bancroft, 1985; Gooren, 1987). It has been variously suggested that the primary behavioral mechanism of action of androgen may involve the enhancement of the rewarding quality of sexual interactions (Davidson, Kwan, and Greenleaf, 1982; Michael and Wilson, 1974), alterations in the threshold for erotic stimulation (Whalen, 1966; Lange, Brown, ’ To whom correspondence should be addressed at Neuropsychiatric Institute and Hospital, Center for the Health Sciences, UCLA 760 Westwood Plaza, Los Angeles, CA 904t241759. 367 0018-506X/91

$1.50

Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.

368

ALEXANDER

AND SHERWIN

Wincze, and Zwick, 1980), or increased sexual interest or motivation (Skakkebaek, Bancroft, Davidson, and Warner, 1980; O’Carroll et al., 1985). The finding that T replacement therapy enhances erectile response to self-generated sexual fantasy in hypogonadal men but is not necessary for erectile response to explicitly erotic films suggests that T may have direct effects on a “cognitive” component of sexual interest (Bancroft and Wu, 1983). For example, sexual fantasy may involve visuospatial abilities or the activation of specific cognitive schemas that are dependent on the presence of T. There may also be a relationship between T levels and attention to erotic stimuli during sexual arousal (Lange et al., 1980; Andrews, 1978). Investigations using objective measures of physiological arousal have found a positive relationship between T and arousability in eugonadal men in response to erotic videotapes (Lange et al., 1980; Rubin, Hensen, Falvo, and High, 1979). On the other hand, men’s self-reports of sexual activities and sexual desire are not related to their levels of T (Raboch and Starka, 1973; Kraemer, Becker, Brodie, Doering, Moos, and Hamburg, 1976; Brown, Monti, and Corriveau, 1978). However, Udry, Billy, Morris, Groff, and Raj (1985) reported that, in adolescent boys, free T level is a better predictor of sexual motivation defined as intentions to have intercourse, reports of sexual arousal and masturbation and the frequency of sexual thoughts, than is physical maturation or age. One explanation for the variable relationship between T and male sexual behavior is that as long as T levels exceed some critical level, higher concentrations of the hormone do not result in greater frequency or intensity of sexual behaviors (Bancroft, 1988). This “threshold” phenomenon is supported by animal research demonstrating that individual differences in copulatory behavior are not a function of individual differences in the amount of circulating androgen (Beach and Holz-Tucker, 1949; Larsson, 1966; Damassa, Smith, Tennent, and Davidson, 1976). The finding that levels of free T predict levels of sexual motivation in adolescent boys (Udry et al., 1985) suggests another explanation for the apparent inconsistencies in the relationship between T and male sexual behavior. It is likely that the expression of sexual interest by adolescent males is less confounded by experiential variables whereas adult behaviors are influenced to a larger extent by learning or habit than by physiological state (Whalen, 1966). Therefore, in sexually experienced adults, selfreports of sexual behavior may be influenced by a combination of biological state, subject expectancies, attitudes, and learning experiences. If this is true, it implies that hormonal effects on sexual motivation in men may be better assessedby measures which are presumably less affected by these nonhormonal factors than are frequencies of sexual activities. It is generally assumed that cognitive aspects of sexual behavior such as sexual thoughts, sexual daydreaming, or levels of sexual desire may pro-

ANDROGEN

AND MALE SEXUALITY

369

vide more sensitive measures of sexual motivation. However, their usefulness is constrained by the fact that self-report measures are not necessarily a reliable reflection of cognitive processes (Nisbett and Wilson, 1977). One method of accessing aspects of cognitive processing independent of self-report is based on the principle of selective interference. In tasks of selective attention, the individual is asked to attend to one dimension of a task and to ignore a distracting element. Typically, when the distracting element contains self-relevant information, performance on selective attention tasks is impaired (Bargh, 1982). An attentional bias toward certain types of distracters has been interpreted as evidence of the activation of existing cognitive structures that code information associated with specific behaviors. For example, research using selective attention paradigms has demonstrated that obsessive-compulsive (Foa and McNally, 1986), phobic (Burgess, Jones, Robertsons, Radclife, and Emerson, 1981), anxious (Mathews and MacLeod, 1986), and premenstrually depressed (McMillan and Pihl, 1987) individuals exhibit an attentional bias for emotionally relevant stimuli that is not apparent in persons without the disorder. This suggested to us that the selective attention model might serve as a heuristic tool to assessmotivational aspects of sexual behavior in eugonadal men that would minimize demand characteristics inherent in self-report measures. Moreover, the salience of sexual stimuli for an individual would likely reflect an aspect of sexuality that is less influenced by learning, experience, and circumstance than are frequencies of sexual behaviors. The goal of this investigation was to examine the relationships between T, sexual attitudes, and cognitive-motivational dimensions of sexual behavior using a dichotic listening task. Men were required to shadow (i.e., repeat aloud) a target message while ignoring a distracting message composed of either neutral or erotic material. It was hypothesized that an attentional bias for sexual stimuli, defined as a greater number of shadowing errors in the presence of the sexual distracter relative to the neutral distracter, would be positively associated with levels of T in eugonadal males. Because of the possibility that cognitive schemas associated with sexual desire would be differentially activated by sexual arousal in men with different levels of T (Lange et al., 1980), the interaction between exposure to a sexual priming stimulus and the subsequent processing of erotic stimuli was also examined. METHOD Subjects Twenty-six heterosexual men between the ages of 18 and 27 responded to recruitment advertisements placed in a university student newspaper.

370

ALEXANDER

AND SHERWIN

Inclusion criteria required that they be single, fluent in English, in good general health and not currently receiving treatment for a psychiatric or physical disorder. Written, informed consent was obtained from all subjects who were paid $10 at the completion of the study. Measures (2) Sexual attitudes and experience. A questionnaire package containing two questionnaires assessedsexual attitudes and experience. The Sexual Experience Scales (SES) (Frenken and Vennix, 1981) describe four dimensions of sexual experience. The Sexual Morality Scale (SES-1) assesses premarital sexual experiences and associated attitudes. The Psychosexual Stimulation Scale (SES-2) assessesthe extent to which an individual seeks or allows sexual stimuli in the form of erotic imagery, the degree of arousal to erotic imagery, and interpersonal sexual attraction. The Sexual Motivation (SES-3) and the Attraction to Marriage (SES-4) Scales describe the sexual and social relationship with a regular partner in terms of the degree of sexual attraction, sexual satisfaction, and the strength of the social and emotional bond to the partner. Subscales derived from the four SES scales provide more specific measures of the various aspects of sexual experience (e.g., arousal by erotic imagery, enjoyment potential of sexual interaction, premarital sexual experience). A general information questionnaire documented demographic information and aspects of the subjects’ sexual history including sexual orientation as assessedby the Kinsey rating scale (Kinsey, Pomeroy, and Martin, 1948), the current frequency of sexual intercourse, number of sexual partners, and marital status. (2) Sexual arousal. Two audiotapes, each approximately 5 min in length, contained the stimuli used to quantify subjective sexual arousal and also served as the priming stimulus for the dichotic listening task. Audiotapes were adopted from a previous study in which it was established that the erotic tape, but not the neutral tape, elicited high levels of subjective sexual arousal in eugonadal male subjects (Hall, 1985). The erotic tape described heterosexual activities including mutual genital fondling, oralgenital contact, and sexual intercourse. It was narrated in the second person by a female who indicated to the subject that he was the sexually active male depicted in the passage. A second audiotape, of equal length and narrated in the same female voice, contained nonsexual material describing a man’s relationship to his family of origin. Quantification of a subject’s response to the audiotape was determined by having him circle a point on a rating scale from 1 (very sexually arousing) to 5 (not at all sexually arousing). (3) Selective attention. A dichotic listening task that required subjects to shadow (i.e., repeat aloud while listening) a message received at one ear (target) while ignoring a message presented simultaneously to the other ear (distracter) was used to measure selective attention for sexual

ANDROGEN

AND MALE SEXUALITY

371

stimuli. Twenty-four targets and 12 neutral distracters were adopted from previous research which had established that their content was neutral in affective tone (Shenker, 1980). Both the targets and neutral distracters were passagesof meaningful prose (120 and 100 words, respectively) with no sexual references or context and provided information on a variety of topics (e.g., wildlife, the weather, and travel). Twelve sexual distracters were chosen from sexual passagestaken from several sources of erotic literature (Lawrence, 1959; Nin, 1977). Each passage consisted of approximately 100 words of meaningful prose describing heterosexual activities including kissing, genital fondling, and oral-genital contact without direct references to male or female genitalia. Eight judges (otherwise not involved in the study) rated the sexual passages and neutral distracters on several descripters (i.e., interesting, aggressive, sexually arousing, offensive) by circling a point on a Likert scale from 1 (not at all) to 7 (extremely). Sexual scripts judged to be the least “offensive” (X = 1.2, ?SEM = 0.05), the least “aggressive” (X = 1.3, aSEM = 0.09), and the least “embarrassing” (X = 1.3, ?SEM = 0.05) were selected to form the sexual distracting element of the dichotic listening task. The 12 sexual distracters were rated as being mildly sexually arousing and significantly more sexually arousing than the 12 neutral distracters (xs = 3.5 ?SEM = 0.14) vs XN = 1.1 +-SEM = 0.12); t(22) = 17.02, P < 0.0001). There were no significant differences between the judges’ ratings of their interest in the material presented in the sexual or neutral distracters (xs = 3.57 (SEM = 0.15) vs XN = 3.54 (SEM = 0.11); t(22) = 0.12, P = 0.90). Targets and distracters were recorded on separate channels of a stereo recorder according to the method outlined by Shenker (1980) and later transferred to audio cassettes. Each target was separated by an interval of 10 set and the onset of each target was indicated by a 1-set 1000~cps tone. Targets were alternately paired with a sexual or neutral distracter whose onset followed the 10th word of the target script. Two forms of the test stimuli (Series 1 and Series 2) were prepared to control for possible target-distracter interactions and order effects. Each series of targetdistracter pairs consisted of two test tapes each containing 12 targetdistracter pairs. Two practice trials and one baseline trial containing target-distracter pairs with only nonsexual descriptive material were also prepared in the same manner as the test stimuli and were then transferred to an audio cassette. The test material was presented on an audio cassette player equipped with stereo headphones. Subjects’ shadowing performances were recorded on audio cassettes by a second audio cassette recorder. (4) Free testosterone. A 20-ml sample of blood was taken by a technician before the practice session. The blood was immediately centrifuged and the plasma stored at -20°C. All samples were assayed in duplicate at

372

ALEXANDER

AND SHERWIN

the conclusion of the study using a Coat-A-Count kit (Diagnostic Products Corp., Los Angeles, CA). The procedure is a direct or single-tube assay whose results are interpolated from a standard curve calibrated in free T concentrations. In this assay, izI-labeled testosterone analog competes for a hxed time with free testosterone in the sample for sites on testosterone-specific antibody immobilized in a polypropylene tube. The antibody-bound fraction is isolated and counted in a gamma counter. The counts are inversely related to the concentration of free T in the sample. The laboratory intra- and interassay coefficients of variation were 7 and 9%) respectively. Procedure

Each subject was tested by a male research assistant on two consecutive days between 2:00 PM and 5:OOPM. This time period for blood sampling was chosen in consideration of the fact that, in men, plasma T levels undergo a diurnal variation (Resko and Eik-Nes, 1966). Subjects signed a consent form stating that each testing session would be approximately 25 min long. The consent form explained that each session would require a blood sample, and that they would be asked to rate their level of subjective sexual arousal to an audiotape and afterward perform a dichotic listening task. The questionnaire package was given to subjects at the end of the first testing session and the completed forms were returned at the second test time. All men were then randomly assigned to a target ear (left ear, IZ = 12, or right ear, n = 14) and a series of the test stimuli (Series 1, n = 13, or Series 2, n = 13). At each test day, following the drawing of a blood sample, subjects were taken to a separate room where they donned earphones. In order to familiarize subjects with the demands of the dichotic listening task, men were asked to perform two practice trials. To reduce possible concerns about performance ability, subjects were told that shadowing errors were common and that we were interested in the types of errors made by groups of men. It was explained to subjects that they were to shadow the messagepresented to the target ear while ignoring the distracting message that would be presented to their other ear. They were further told that the messagethey were to repeat would be preceded by a tone and would always begin before the distracting message. Following the two practice trials, subjects were presented with one baseline trial. The subjects were told that their performance during the baseline and test session would be recorded and scored for shadowing errors at the completion of the investigation. Following the baseline trial, subjects were presented with either the 5min erotic or neutral priming audiotape. The order of the presentation of the erotic or neutral prime was counterbalanced across the two sessions. After listening to the tape, subjects completed the sexual arousal rating scale which was deposited in a sealed box where it remained until the

ANDROGEN

AND MALE SEXUALITY

373

BLOOD SAMPLE I

I PRACTICE TRIALS (2) BASELINE TRIAL

PrimtiDcl Stimulus

PrW&iUmUs EROTIC AUDIOTAPE SESSION 1 (n=14) SESSION 2 (n- 12)

NEUTRAL AUDIOTAPE SESSION I (n-12) SESSION 2 (n-1 4)

SEXUAL AROUSAL RATING ( I-S SCALE)

TEST TRIALS presentation 01 the 12 target-distracter pairs

FIG. 1. Experimental design.

completion of the investigation. They were then presented with one of the two test stimuli tapes containing the 12 target-distracter pairs. A diagram of the experimental design appears in Fig. 1. Statistical Analyses

Questionnaire items were analyzed using Student t and x2 tests where appropriate. Analyses of variance (ANOVA) with repeated measures was used to assessperformance on the dichotic listening task. Newman-Keul’s post-hoc tests were used to analyze significant interactions. Because of the exploratory nature of the investigation, a stepwise multiple regression was used to assessthe relationships between sexual attitudes, experience, free T, sexual arousal, and selective attention for erotic stimuli. RESULTS Subject Characteristics

Table 1 summarizes the results of the questionnaire package. The average rating on the Kinsey sexual orientation scale indicated that subjects

374

ALEXANDER

AND SHERWIN

TABLE 1 Characteristics of Male Volunteers (n = 26) Mean (*SD)

Age

Kinsey rating scale Postsecondary education (years) SES-1 SES-2 SES-3” SES-4 Free T (pm/l) Session 1 Session 2 Religious affiliation Jewish Christian Other Agnostic No response Marital Status Unmarried/no regular partner Unmarried/regular partner Cohabiting Frequency of intercourse (past year) Never Not within the past 12 months l-10 times/year l-4 times/month 2-4 times/week 5 or more times/week Number of past sexual partners None l-2 partners 3-6 partners 7-10 partners 11-14 partners >15 partners

21.0 (k2.02) 1.1 (kO.4) 4.2 (41.8) - 0.49 (I 0.57) - 0.89 ( -c0.63) -0.20 (kO.19) - 0.67 ( I 0.48) 73.9 (kO.12.4) 73.4 (&0.14.0) n

1 2 6 3 12 2 1 11 8 4 1 1

* Mean of only men with regular sexual partners (n = 15).

were almost exclusively heterosexual. Analyses of the SES subscales that assessedorgasmic response (OE, orgasm adequacy during intercourse and its evaluation) and sexual satisfaction (EP, enjoyment potential of sexual interaction) found that all scores fell within the normal male range. Free T

All subjects had levels of free T well within the normal range of male values (28-156 pmole/liter). There was a significant positive relationship

ANDROGEN

AND

MALE

SEXUALITY

375

between levels of free T at the two sampling times (r(n = 26) = 0.53, P < 0.01). No significant correlations between free T levels and the SES or general questionnaire measures were apparent. Sexual Arousal

Ratings on the 5-point scale that evaluated sexual arousal to the 5-min audiotapes confirmed that the erotic audiotape was more sexually arousing than the neutral tape (xs = 2.23, +SEM = 0.18 vs XN = 4.88, +SEM = 0.08: t(24) = - 12.8, P < 0.001). Whereas 14 men indicated high sexual arousal to the erotic tape (ratings of 1 or 2), the remaining 12 individuals rated the erotic tape as being only mildly sexually arousing (ratings of 3 or 4). There was no relationship between the arousal group and the order in which men listened to the priming audiotape (x” = 0.87, P > 0.05). The difference between the average sexual arousal ratings in the high sexual arousal (HSA) and low sexual arousal (LSA) groups was statistically significant (xu = 1.57 +SEM = 0.13 vs xL = 3.31 &SEM = 0.18: t(24) = -7.81, P < 0.001). On the other hand, the HSA and LSA groups could not be discriminated on the bases of age, years of schooling, intercourse frequency, strength of heterosexual orientation rating, or numbers of past sexual partners (eigenvalue = 0.07, P = 0.93). A stepwise multiple regression using sexual arousal ratings as the dependent variable was performed using average free T and SES scales as predicter variables. The only variable to enter the model was the arousal to erotic imagery (AE) subscale (F(1, 24) = 6.3, P < 0.05). Scores on this scale accounted for 21% of the variance in sexual arousal ratings. A more negative score on this scale indicates higher levels of sexual arousal to erotic imagery. Thus, HSA raters usually experienced higher levels of sexual arousal to erotic imagery than did the LSA raters (xu = - 1.76, + SEM = 0.21 vs -0.90, +SEM = 0.22; t(24) = -2.82, P < 0.01). Although there X L= was a trend for higher levels of free T to be associated with higher levels of sexual arousal (r(26) = -0.29, P < O.lO), average levels of T did not differ between HSA and LSA groups (xu = 76.02, +SEM = 3.5 and = 71.3, ?SEM = 2.78; t(24) = 1.05, ns). XL Shadowing Performance

Shadowing performances were assessed using an established scoring procedure (Shenker, 1980; McMillan and Pihl, 1987). Four scores representing the average number of errors in the presence of the sexual distracter and the average number of errors in the presence of the neutral distracter for each priming condition were obtained for each subject. Table 2 presents a summary of the men’s performance on the dichotic listening task at baseline and after exposure to the erotic or neutral prime. There were no significant differences between the mean number of shadowing errors made by HSA and LSA groups during the two baseline trials. Men

376 Mean (*SD)

ALEXANDER

AND SHERWIN

TABLE 2 Number of Shadowing Errors Made by Heterosexual Men in the Presence of Sexual and Nonsexual Distracters High arousal raters* n = 14

Baseline: session 1 Baseline: session 2

7.2 (k1.8) 12.4 ( k 2.4)

Low arousal raters? n = 12 9.8 (24.2) 10.9 (23.1)

Neutral distracter Sexual distracter

Performance after erotic prime 8.2 (k5.8) 13.9 (k9.7)b

9.3 (-+7.3) 15.0 (+9.1)b

Neutral distracter Sexual distracter

Performance after neutral audiotape 10.0 (k5.7) 18.0 (2 10.3)*

8.7 (e7.3) 13.4 (+7.4)b

* A sexual arousal rating of 1 or 2 on a 5-point scale. t A sexual arousal rating of 3 or 4 on a 5-point scale. b > a, P < .OOl.

in both groups made significantly more shadowing errors across trials when the distracter contained sexual material than they did when the distracter contained neutral material (x, = 15.1 ? SEM = 0.59 vs xN = 9.1 +SEM = 0.5; F(l, 24) = 20.87, P < 0.001). (a) Factors associated with shadowing accuracy. There was a significant interaction between the effects of sexual arousal rating group (HSA or LSA) and priming condition on shadowing accuracy (F(l, 24) = 4.43, P < 0.05). Following exposure to the neutral prime, men in the HSA group made a greater number of shadowing errors than those in the LSA group (P < 0.05). There were no differences between the HSA and LSA groups in the number of shadowing errors made following the erotic priming audiotape. Within-group analyses indicated that, in the HSA group, shadowing errors were less frequent following exposure to the erotic priming stimulus relative to performance after exposure to the neutral priming stimulus (P < 0.05) (Fig. 2). Average levels of free T were significantly and negatively correlated with task performance following the erotic prime, such that higher levels of free T were associated with fewer shadowing errors (r(n = 26) = -0.39, P c 0.05). (b) Factors associated with attention to sexual stimuli. Multiple regression analyses assessedthe relationship between the frequency of errors made in the presence of the sexual distracter (ESD) and average free T, SES1, and SES-2 scale scores. To control for individual differences in task ability, subjects’ average baseline performance scores were forced into the regression model in the first step. This variable accounted for 25% of the variance in ESD (F(l, 24) = 8.06, P < 0.01). SES scales and hormone levels were added to the model in a stepwise manner. Only the

ANDROGEN

AND MALE SEXUALITY

317

63 High sexual arousal group 0 LOWsexual arouml group

5 %

5

neutral Priming Condition 2. The mean frequency (+SEM) of shadowing errors in high arousal raters (n = 14) and low arousal raters (n = 12) after exposure to the erotic and neutral priming audiotape. (Note: means with same letters are not significantly different.) FIG.

AE subscale of the SES-2 met the criteria for entry into the model (F(2, 23) = 8.08, P < 0.01). Higher levels of arousal to erotic imagery were correlated with greater ESD scores (r = -0.46, P < 0.01). Together, the average baseline performance and AE scores accounted for 43% of the variance in ESD scores. A second stepwise multiple regression was performed using subject’s average bias toward the sexual distracter as the dependent variable. This score was the difference between the number of errors made in the presence of the sexual and neutral distracters (bias = sexual - neutral). Average free T levels and SES scales were used as independent variables. Again, only the AE subscale of the SES entered into the model (F(1, 24) = 5.5, P < 0.05) and accounted for 19% of the variance in bias scores. Higher levels of arousal to erotic imagery were again associated with a greater bias toward the sexual distracter. DISCUSSION The novel aspect of this investigation was the application of a methodology used in experimental cognitive research to the study of hormones and attention to sexual information in young, healthy, heterosexual men. Attention may be consciously or unconsciously allocated to sexual stimuli in the environment. For example, during sexual activities individuals often direct their attention toward relevant sexual cues. It has also been suggested that the diversion of attention to the presence of sexual stimuli from a conscious focus on nonsexual stimuli is an indicator of normal

378

ALEXANDER

AND

SHERWIN

levels of sexual desire (Schreiner-Engel, Schiavi, White, and Ghizzani, 1989). Although individuals likely differ in terms of the strength of the tendency to be distracted by sexual cues, it is not known whether there is any relationship between sex steroids, sexual attitudes, sexual behavior, and a selective attention bias for sexual stimuli. Contrary to our hypothesis, there were no positive correlations between levels of free T and individual differences in the selective attention bias for sexual stimuli. However, the strength of the attentional bias for sexual stimuli in young, healthy, men was found to be strongly and positively associated with men’s reports of their usual levels of sexual arousal to erotic imagery. This finding is noteworthy given the evidence that conscious attentional focus on sexual cues enhances level of sexual arousal in men (Geer and Fuhr, 1976; Dekker and Everaerd, 1989). Importantly, there was good consistency between the questionnaire measure of sexual arousal and men’s response to the erotic priming stimulus. Thus, whether allocation of attention to sexual stimuli is conscious or unconscious, the degree to which sexual stimuli may activate those schemas that code information on sexual stimuli and sexual response appears to be positively related to the propensity for sexual arousal in men. The wide range of men’s self-reports of sexual arousal to the erotic audiotape permitted a further assessmentof the relationship between the degree of subjective sexual arousal and task performance. When men were divided into high and low arousal groups based on their ratings of the erotic priming audiotape, neither age, sexual experience, sexual attitudes, nor average levels of free T discriminated between them. Consistent with the finding that high sexual arousal was associated with a greater salience of the sexual stimuli presented during the dichotic listening task, men in the high arousal group made more shadowing errors following exposure to the neutral prime than men in the low sexual arousal group. While the high sexual arousal group continued to exhibit a readiness to attend to sexual stimuli, their shadowing performance improved following the erotic priming audiotape relative to their performance following the neutral prime. Together, these results suggest not only that men with lower thresholds for sexual arousal attend to sexual information to a greater degree than men with higher thresholds for sexual arousal, but also that arousal may enhance their general ability to consciously direct and maintain attention. Irrespective of the level of sexual arousal to the erotic tape, the ability to consciously direct attention to the experimentally defined relevant task dimension (target) was associated with free T such that men with higher levels of T made fewer errors following exposure to the erotic prime. This is consistent with the hypothesis that T may act centrally to enhance persistence of attention to relevant cues (Andrew, 1978). However, this explanation fails to account for the finding that the T-attention association

ANDROGEN

AND MALE SEXUALITY

379

was apparent following exposure to the erotic prime but not following exposure to the neutral prime. Keeping in mind that shadowing errors were fewer in men with higher levels T following the erotic prime, the fact that men with higher levels of sexual arousal also made fewer errors following the erotic prime suggests there may be a relationship between T, sexual arousal, and attention in young, healthy men. This possibility gains tentative support from the observed trend for a positive relationship between T and sexual arousal ratings of the erotic priming audiotape. However, a study using a larger number of subjects or a measure of sexual arousal more sensitive than a 5point rating scale is needed to clarify the association between T and sexual arousal. It has been suggested that arousal and arousability reflect levels of sexual motivation more accurately than levels of sexual fantasy, masturbation frequency, or sexual interactions with partners (Whalen, 1966). Arousal to sexual imagery is thought to require the activation of an associative network of propositions in memory which must include those propositions describing the emotional response (Lang, 1979; Dekker and Everaerd, 1988). The positive relationships documented in this investigation between sexual arousal and sensitivity to sexual stimuli and between T and attention to relevant stimuli suggest that T may play a role in the activation of sexual response propositions that exist in memory. A comparison of the processing of sexual information in eugonadal and hypogonadal men would constitute an important test of this hypothesis and may lead to a better understanding of the role of T in sexual arousal and the processing of sexual information. ,ACKNOWLEDGMENTS This work was supported by Grant MA-8707 from the Medical Research Council of Canada awarded to B. B. Sherwin. The authors thank Walter Marcantonio and Daniel Weiner for assistance in the collection of the behavioral data and Naomi Epstein for drawing blood samples.

REFERENCES Andrew, R. J. (1978). Increased persistence of attention produced by testosterone and its implications for the study of sexual behavior. In J. Hutchison (Ed.), Biological Determinants of Sexual Behavior, p. 255. Wiley, Chichester. Bancroft, J. (1988). Sexual desire and the brain. Sex Marital Ther. 3, 11-27. Bancroft, J., and Wu, F. C. W. (1983). Changes in erectile responsiveness during androgen replacement therapy. Arch. Sex. Behav. 12, 59-66. Bargh, J. A. (1982). Attention and automaticity in the processing of self-relevant information. J. Person. Sot. Psych. 43, 425-436. Beach, F. A., and Holtz-Tucker, A. M. (1949). Effects of different concentrations of androgen upon sexual behavior in castrated male rats. J. Comp. Physiol. Psych. 42, 433-445. Brown, W. A., Monti, P. M., and Corriveau, D. P. (1978). Serum testosterone and sexual activity and interest in men. Arch. Sex. Behav. 7, 97-103.

380

ALEXANDER

AND SHERWIN

Burgess, I., Jones, L., Robertsons, S., Radclife, W., and Emerson, E. (1981). The degree of control exerted by phobic and non-phobic verbal stimuli over the recognition behavior of phobic and non-phobic subjects. Behav. Res. ‘J’her. 19, 233-243. Damassa, D. A., Smith, E. R., Tennent, B., and Davidson, J. M. (1976). The relationship between circulating testosterone levels and male sexual behavior in rats. Hot-m. Behnv. 8, 275-286. Davidson, J. M., Carmargo, C. A., and Smith, E. R. (1979). Effects of androgens on sexual behavior of hypogonadal men. J. Clin. Endocrinol. Metab. 48, 955-958. Davidson, J. M., Kwan, M., and Greenleaf, W. J. (1982). Hormonal replacement and sexuality in men. Clin. Endocrinol. Metub. 11, 599-623. Dekker, J., and Everaerd, W. (1988). Attentional effects on sexual arousal. Sot. Psychophysiol.

Res. 25, 45-54.

Dekker, J., and Everaerd, W. (1989). A study suggesting two kinds of information processing of the sexual response. Arch. Sex. Behav. 18, 435-447. Foa, E. B., and McNally, R. J. (1986). Sensitivity to feared stimuli in obsessive-compulsives: A dichotic listening analysis. Cog. Ther. Res. 10, 477-485. Frenken, J., and Vennix, P. (1981). Sexuality Experience Scales. Swets & Z.&linger B. V., Zeist, The Netherlands. Geer, J. H., and Fuhr, R. (1976). Cognitive factors in sexual arousal: The role of distraction. J. Consult. Clin. Psychol. 44, 238-243.

Gooren, L. J. G. (1987). Androgen levels and sex functions in testosterone-treated hypogonadal men. Arch. Sex. Behav. 16, 463-473. Hall, K. S. K. (1985). Alcohol and Male Sexual Arousal: The Effects of Rising and Falling Blood Alcohol Levels. Unpublished doctoral dissertation, McGill University, Montreal. Kinsey, A. C., Pomeroy, W. B., and Martin, C. E. (1948). Sexual Behavior in the Human Male. Saunders, Philadelphia. Kraemer, H. C., Becker, H. B., Brodie, H. K. H., Doering, C. H., Moos, R. H., and Hamburg, D. A. (1976). Orgasmic frequency and plasma testosterone levels in normal human males. Arch. Sex. Behav. 5, 125-132. Kwan, M., Greenleaf, W. J., Mann, J., Crapo, L., and Davidson, J. M. (1983). The nature of androgen action on male sexuality: A combined laboratory-self-report study on hypogonadal men. J. Clin. Endocrbtol. Metab. 57, 557-562. Lang, P. J. (1979). A bio-informational theory of emotional imagery. Psychophysiology,

16, 495-512. Lange, J. D., Brown, W. A., Wincze, J. P., and Zwick, W. (1980). Serum testosterone concentration and penile tumescence changes in men. Horm. Behav. 14, 267-270. Larsson, K. (1966). Individual differences in reactivity to androgen in male rats. Physiol. Behav.

1,255-2X

Lawrence, D. H. (1959). Lady Chatterley’s Lover. Grove Press, New York. (Original work published in 1928.) Mathews and MacLeod (1986). Discrimination of threat cues without awareness in anxiety states. J. Abn. Psych. 95, 131-138. McMillan, M. J., and Pihl, R. 0. (1987). Premenstrual depression: A distinct entity. 1. Abn. Psych. 96, 149-154. Michael, R. P., and Wilson, M. (1974). Effects of castration and hormone replacement in fully adult male rhesus monkeys (Mucucu muluttu). Endocrinology 95, 150-159. Nin, A. (1977). Delta of Venus. Harcourt Brace Javanovick, New York. Nisbett, R. E., and Wilson, T. D. (1977). Telling more than we can know: Verbal reports on mental processes. Psych. Rev. 84, 231-259. O’Carroll, R., Shapiro, C., and Bancroft, J. (1985). Androgens, behaviour, and nocturnal erection in hypogonadal men: The effects of varying the replacement dose. Clin. Endocrinol.

23, 527-538.

ANDROGEN

AND MALE SEXUALITY

381

Raboch, J., and Starka, L. (1973). Reported coital activity of men and levels of plasma testosterone. Arch. Sex. Behav. 2, 309-315. Resko, J. A., and Eik-Nes, K. B. (1966). Diurnal testosterone levels in peripheral plasma of human male subjects. J. Clin. Endocrinol. 26, 573-576. Rubin, H. B., Henson, D. E., Falvo, R. E., and High, R. W. (1979). The relationship between men’s endogenous levels of testosterone and their penile responses to erotic stimuli. Behav. Res. Ther. 17, 305-312. Schreiner-Engel, P., Schiavi, R. C., White, D., and Ghizani, A. (1989). Low sexual desire in women: The role of reproductive hormones. Horm. Behav. 23, 221-234. Shenker, L. J. (1980). Selective Attention to Dysphoric Stimuli by Depressed and Nondepressed Individuals. Unpublished doctoral dissertation, McGill University, Montreal. Skakkeback, N. E., Bancroft, J., Davidson, D. N., and Warner, P. (1981). Androgen replacement with oral testosterone undecanoate in hypogonadal men: A double blind controlled study. Clin. Endocrinol. 14, 49-61. Udry, J. R., Billy, J. 0. G., Morris, N. M., Groff, T. R., and Raj, M. H. (1985). Serum androgenic hormones motivate sexual behavior in adolescent boys. Fertil. Steril. 43, 90-94. Whalen, R. E. (1966). Sexual motivation. Psych. Rev. 73, 151-163.