Folia Parasitologica 41: 122-126, 1994.

J. FLEGR, I. HRDÝ

Department of Parasitology, Faculty of Science, Charles University, Vinicna 7, Prague 128 44, Czech Republic

Running title: Effect of toxoplasmosis on a human personality

Key words: Toxoplasma, toxoplasmosis, behavior, personality factors, Cattell's questionnaire, manipulation hypothesis.

Abstract. An effect of parasites on host behavior was tested on the toxoplasma-human model. Three hundred and thirty-eight (338) people were assessed with Cattell's personality questionnaire and then tested for Toxoplasma gondii infection with a delayed type hypersensitivity test for Toxoplasma. A highly significant correlation between chronic toxoplasmosis and two personality factors (G- Low Superego Strength and L- Protension) was found (p=0.0032 and 0.0020, respectively). A correlation of the intensity of the personality factor-shifts with the duration of the infection (estimated from antibody titer) suggests that toxoplasmosis induces the shift in human personality, rather than the personality factors G and L influence an acquisition rate of Toxoplasma gondii infection.

Correspondence to: Jaroslav Flegr, Department of Parasitology, Faculty of Science,                                     Charles University, Vinicna 7, Prague 128 44, Czech Republic.

The effect of parasitism on host behavior is being demonstrated in a growing number of host-parasite systems. The induced behavioral patterns often promote transmission of the parasite. One hypothesis suggests that such modification of host behavior is a sophisticated product of parasite evolution aimed at host manipulation rather than an accidental byproduct of other physiological activities of the parasite (Barnard and Behnke 1990, Dawkins 1982).

Among parasitic protozoa from the suborder Eimeriina, the phenomenon has been observed in Sarcocystis (Hoogenboom and Dijkstra 1987), Eimeria (Kavaliers and Collwell 1982) and Toxoplasma gondii Nicolle et Manceaux 1906 (Hay et al.1985). The last organism is an intestinal coccidian of felids with an unusually wide range of intermediate hosts, including humans. After an acute phase of infection (promoted by tachyzoites), in Toxoplasma-infected subjects cysts are formed mainly in neural and muscular tissues. The parasites (bradyzoites) inside the cysts usually cause little or no harm in immunocompetent individuals and probably persist as viable parasites for the life of the host (Remington and Krahenbuhl 1982). The transmission of the parasite from the intermediate to the definitive host is mediated by carnivorism. Since sexual reproduction of Toxoplasma can be accomplished only in the feline, there might be a strong selection pressure on the parasite to evolve a mechanism to manipulate intermediate host-behavior so as to favor predation by felines. A high prevalence of lifelong toxoplasmosis (about 32% of pregnant women in New York City, 22% in London and 84% in Paris (Desmonts and Couvreur 1974)) offers an opportunity to study the possible influence of the parasitism on human behavior by random screening of normal population. Here we studied an effect of chronic toxoplasmosis on human personality in an experimental set of 338 humans.

MATERIAL AND METHODS

Subjects

The study was carried out in the Faculty of Science, Charles University, Prague. Data were collected for a period of 14 months (1992-1993). The experimental set contained 195 men and 143 women, mostly zoological departments staff and biology students. The main advantage of this experimental set is its relative homogeneity. On the other hand the extrapolation of our results on different sets should be done carefully. All subjects gave their informed consent before they were accepted for the study.

Personality test

Cattell's sixteen factor questionnaire (form A) (Cattell 1970) was used for the characterization of personalities. This questionnaire is widely used in personality studies in the Czech Republic. It covers sixteen personality factors (Tab. 1). The main advantage of this test is that it contains only one hundred and eighty-seven (187) questions. Therefore, it can be completed by most subjects within one hour. With the exception of parasitologists, practically all subjects were tested psychologically before the results of the toxoplasmosis test were known.

Immunological test for chronic toxoplasmosis

As the chronic toxoplasmosis we understand here the presence of anti-Toxoplasma immunity in human subjects without any clinical symptoms of acute toxoplasmosis. The existence of specific immunity was assessed by intradermal delayed hypersensitivity test (IDHT) (Feldman 1954). The test was performed using toxoplasmin (SEVAC, USOL Prague) as the antigen and shame injection of pure solute as the negative control. Positive reactions were of the delayed tuberculin type and were measured at 48 h following antigen administration. The large-scale use of the toxoplasmin skin test in population surveys has yielded excellent agreement between the results of this test and the presence or absence of humoral antibody (Remington and Krahenbuhl 1982). The ability to elicit DH to toxoplasma antigens in man appears to require months to years after initial infection to develop, so it appears to be most useful in the diagnosis of chronic infection (Remington and Krahenbuhl 1982).

In those persons giving their consent (41 subjects), the presence of anti-Toxoplasma antibodies was evaluated by an indirect fluorescent antibody test, IFAT, (Goldman 1957) using the IFR kit prepared in our lab (Kramař et al. 1963). There was a perfect agreement between the results of IDTH and IFAT.

Statistics

The BMDP (Dixon 1985) program was used for all of the following statistical testing: The Hotelling t-test was used for comparing the sets of Toxoplasma-infected and Toxoplasma free subjects (P3D). A Two-way analysis of variance ANOVA (P7D) was used to study the effects of toxoplasmosis, age and toxoplasmosis-age interactions on personality factors. Discriminant analysis (P7M) was used for personality factor-based diagnosis of toxoplasmosis and for stepwise covariance analysis of personality factors. Kendall nonparametric range correlation test (P3S) was used to estimate the correlation between the anti-Toxoplasma antibodies level and the amount of personality factors-shift.

RESULTS

Personality profiles of two groups of subjects, Toxoplasma infected ones (TI) (56 men and 34 women) and Toxoplasma free ones (TF) (139 men and 109 women) were compared using the Hotelling t-test. No difference in personality profiles between TI and TF groups was detected when women and men were considered together (p=0.241) or when only TI and TF women were compared (p=0.204). However, a highly significant difference was detected when the personality profiles of TI and TF men were compared (p=0.025). Of the 16 factors estimated by the Cattell's questionnaire the four factors which apparently caused the difference are marked in Table 1. Because of this result, all the following testing were performed only with the men set.

Table 1.

The names and characteristics in the left column hold for persons with low value of the factor, those in the right columns for persons with the high value of the factor. The raw data for any person were standardized with respect to his/her age and sex. The single star (p<0.05) and two stars (p<0.01) designate that property that predominated among the toxoplasma-positive people.

SIZOTHYMIA *	A	AFFECTOTHYMIA
reserved, detached, critical		warm-hearted, outgoing, easygoing
LOW INTELLIGENCE	B	HIGH INTELLIGENCE
EGO WEAKNESS	C	HIGH EGO STRENGTH
affected by feelings, emotionally less stable		stable, mature, faces reality, calm
SUBMISSIVENESS	E	DOMINANCE OR ASCENDANCE
obedient, mild, easily led, docile		aggressive, competitive, stubborn
DESURGENCY	F	SURGENCY
sober, taciturn, serious		enthusiastic, heedless, happy- go -lucky
LOW SUPEREGO STRENGTH **	G	SUPEREGO STRENGTH, CHARACTER
disregards rules, expedient		conscientious, persistent, moralistic, staid
THRECTIA	H	PARMIA
shy, timid, restrained,		adventurous, "thick-skinned", socially bold
HARRIA	I	PREMSIA
tough-minded, rejects illusions		tender-minded, sensitive
ALAXIA	L	PROTENSION **
trusting, accepting conditions, tolerant		suspecting, jealous, dogmatic
PRAXERNIA	M	AUTIA
practical, has "down to earth"		imaginative, bohemian
NAIVETÉ	N	SHREWDNESS
forthright, unpretentious		astute, worldly, polished
UNTROUBLED ADEQUACY	O	GUILT PRONENESS
self-assured, placed, secured		apprehensive, self-reproaching, insecure
CONSERVATISM OF TEMPERAMENT	Q1	RADICALISM
conservative, respecting		experimenting, liberal, analytical
GROUP DEPENDENCY	Q2	SELF-SUFFICIENCY
sociably group dependent, "joiner"		self-sufficient, resourceful, prefers own decisions
LOW SELF-SENTIMENT INTEGRATION*	Q3	HIGH STRENGTH OF SELF SENTIMENT
uncontrolled, lax, follows own urges		controlled, exacting will power, socially precise
LOW ERGIC TENSION	Q4	HIGH ERGIC TENSION
relaxed, tranquil, torpid		tense, frustrated, driven, overwrought

Some personality factors as well as the frequency of toxoplasmosis vary with the age of a subject. To determine whether the differences observed were the result of causal relationships between toxoplasmosis and the personality factors or whether it was only a false correlation resulting from the effect of age, a Two-way ANOVA was performed. This analysis showed that for factors G and L and possibly A and Q₃ as well, the toxoplasmosis not age was responsible for the observed differences (Fig 1).

Fig. 1.

Effect of toxoplasmosis and age on four personality factors estimated in 195 male subjects by Two-way analysis of variance (BMDP, program P7D). The effects of toxoplasmosis, age and toxoplasmosis-age interactions were tested. Only the tail probabilities for toxoplasmosis effects are given. Open and crosshatched bars show the means with standard errors of a particular personality factor for different age classes of Toxoplasma negative and positive subjects, respectively. The cut points for the age classes correspond to those used for the age standardization of row data obtained by Cattell's questionnaire. The numbers of subjects in groups on the graph were indicated on the upper plot. The equality of variances for the factors listed was confirmed by the Levane's test.

The possibility that some of the personality factors changed because of their correlation with another toxoplasmosis-influenced factor rather than their correlation with toxoplasmosis was studied by a discriminant analysis. This statistical method was primarily designed for the classification of objects into previously defined groups. It can also be used for estimating the correlation effects among the factors by an approach based on analysis of covariance (Bouška et al.1990). During a stepwise process of incorporating factors A, G and L into a classification function, it appears that factor A (Sizothymia) was strongly influenced by factor L (Protension). By incorporating the factor L, the F value for factor A decreased from 2.90 to 2.40, while by incorporating factor A, the F value for L decreased from 7.37 to 6.84. Factors G and L, however, were entirely independent (after incorporation L, the F value for G only decreased from 6.91 to 6.90).

Discriminant analysis performed on the basis of Cattell´s personality factors proved to be able to sort out group of subjects suspected of having toxoplasmosis. In the group of 14 subjects classified as toxoplasma positive on the basis of the factors G and L, the frequency of toxoplasmosis was 64% (a priori probability, the frequency in the unsorted population was 28.7%). When factors F and Q₂ were included into classification function, the frequency of correctly diagnosed toxoplasmosis increased up to 73.7%.

Among 176 subjects classified as Toxoplasma negative the frequency of correct diagnosis was 76.1%. That differs only slightly from a priori probability 71.3%. Such results could be expected if the Toxoplasma infection antecedes the shift in the personality factors. While a relatively recently acquired toxoplasmosis (months) can be effectively detected by immunological test (Krahenbuhl and Remington 1982), the transformation of the human personality (and the manifestations of the changes) evidently could be a long term process. This offers a possibility to test whether the toxoplasmosis induces the personality factor-shift or whether some personality factors influence the probability of being infected by T. gondii. Early infections can be recognized by higher level of specific antibodies (Krahenbuhl and Remington 1982). For 24 infected male subjects the data from indirect fluorescent antibody test (IFAT) were available. Using a nonparametric Kendall rank correlation test, a highly significant correlation (k= -0.573, p=0.00009) occurred between the levels of the antibodies measured by IFAT and the amount of personality factors-shifts, quantified with the values of posterior probabilities for group TI obtained by discriminant analysis, Fig. 2. The existence of such correlation suggests that the personality factor-shift develop after the T. gondii infection.

Fig. 2.

Correlation of the level of anti-Toxoplasma antibodies with the amount of personality factor shift. Each point corresponds to a single Toxoplasma infected subject. x-axis: antibody titer measured by IFAT, y-axis: posterior probability of having toxoplasmosis computed on the basis of Cattell's personality factors by discriminant analysis. The value of Kendall rank correlation coefficient, k  =  -0.573 confirms the existence of the correlation on the level p = 0.00009.

DISCUSSION

Our results suggest that in human males a strong correlation exists between certain personality factors and chronic toxoplasmosis (monitored by IDTH). The nature of critical personality factors (see Tab. 1) and a positive correlation between the amount of personality factors-shift and the antibody titer-based estimates of the duration of the infection suggests that the toxoplasmosis induces the shift in human personality, rather than the personality factors influence an acquisition rate of Toxoplasma gondii infection.

The positive reaction in IDHT suggests the existence of cellular anti-Toxoplasma immunity. The presence of the immunity is mostly interpreted to be a marker of acute or chronic toxoplasmosis (Krahenbuhl and Remington 1982). Experience with AIDS patients suggests that the frequency of latent human toxoplasmosis may be quite similar to the frequency of Toxoplasma-immune subjects. However, it still must be clarified as to whether the personality factors-shift was induced by latent chronic toxoplasmosis or whether it is only triggered by the acute stage of the infection.

The results from the ANOVA and from t-tests suggest that personality factors G and L are, and the A and Q₃ might be, shifted in TI persons. On the other hand, in discriminant analysis another group of factors (G, L, F and Q₂) proved to be most suitable for identification of TI subjects. This discrepancy can be explained by the differences in the mathematical backgrounds of these two methods. Discriminant analysis includes a stepwise covariance analysis. In every step, the computed F-to-enter values are affected by the variables already present in the classification function. When a strong correlation exists between two variables (e.g., factor A and L) only the one with the higher F value might be entered classification function. On the other hand, when there is no correlation among variables (e.g., factors G, L, F and Q₂), all might be entered the function, no matter what their F values in the step zero of discriminat analysis (or in t-tests.) Our results suggest that toxoplasmosis influenced the factors G, L, F and Q₂. The factors A and Q₃ might also be shifted in TI male subjects. Their shifts, however, were caused be their correlation with factors G or L, rather then with toxoplasmosis.

The personality factors-shift was demonstrated only in men. It is possible that women are more resistant to manipulation by the parasite. Another explanation is that our technique of evaluating human personality factors should be modified for the female population. It is well known (Steiner 1987, Mohan and Chopra 1986) that results obtained with a personality questionnaire in women vary when the tests are performed in different stages of the menstrual cycle. It can be assumed that similar set of personality markers (biologically based ones) could be influenced by hormone levels as well as by toxoplasmosis. Therefore, it will be important to perform further studies in women through all menstrual cycle stages and to collect the data from all experimental subjects in the same stage.

The influence of severe forms of acute toxoplasmosis on human personality has been reported by clinicians (Minto and Roberts 1959, Burkinshaw et al.1953, Freytag and Haas 1979, Ladee et al.1966). Typically, however, the effects of acute toxoplasmosis on immunocompetent patients are mild. Chronic toxoplasmosis is usually considered asymptomatic (Remington 1974). In fact, only a negligible fraction of infected people ever learn that they are parasitized by Toxoplasma gondii. It seems reasonable to expect that the observed psychological symptoms of toxoplasmosis only indirectly reflect some activities of the parasite. One can only speculate, as to whether these activities are connected with an attempt (in the human host a nonproductive one) of the parasite to manipulate the host behavior, or are only byproducts of a nonspecific worsening of the quality of the host life. Theoretically, infected subjects might suffer more frequent or more severe diseases because of an interference of the parasite with their immune system (Remington and Krahenbuhl 1982). Such effects of toxoplasmosis, however, have never been reported.

The apparent influence of T. gondii on the personality of human males was unexpected and pronounced. The nature of the shift in human personality following Toxoplasma infection can hardly be considered desirable. It might seem anecdotal that in a group of 29 Toxoplasma negative professors from our experimental set there are 10 present or past decision-makers (heads of department, vicedeans and deans) while among 14 Toxoplasma positive professors there is only a single head of department.

Acknowledgements.

This research was supported by the Czech Biological Foundation. We thank A. Exnerova for her help with psychological questionnaires, J Chalupský and D. Eremiáš for performing IFAT and V. Albrecht, J. Zocová, H. Maslowská, M. Černý and J. Hrdá for the statistical analysis. We also thank our colleagues that voluntarily participated in the experiments.

REFERENCES

BARNARD C.J., BEHNKE J.M. (Eds.) 1990: Parasitism and Host Behaviour. Taylor and Francis, London, New York, Philadelphia, 332 pp.

BOUŠKA V., DELANO J. W., MASLOWSKÁ H., ŘANDA Z. 1990: Discriminant analysis of the moldavite data set. Acta Universitatis Carolinae-Geologica 3: 243-258.

BURKINSHAW J., KIRMAN B. H., SORSBY A. 1953: Toxoplasmosis in relation to mental deficiency. Brit. Med. J. 702: 1-6.

CATTELL R.B. 1970: Handbook for the sixteen personality factors questionnaire (16PF). Institute for Personality and Ability Testing, Champaign.

DAWKINS R. 1982: The Extended Phenotype: the Gene as the Unit of Selection. Freeman, London.

DESMONTS G., COUVREUR J. 1974: Congenital toxoplasmosis. A prospective study of 378 pregnancies. N. Eng. J. Med. 290: 1110-1116.

DIXON W. J. 1985: BMDP Statistical Software. Univ. Calif. Press, Berkeley.

FELDMAN H. A. 1954: Laboratory methods in current use for the study of toxoplasmosis. Advances in Ophthalmology Vol. III, Streiff E. B., (Ed.), Lausanne Bibliotheca Ophthalmologica Fasc. 39: 1-11.

FREYTAG H. W., HAAS H. 1979: Zur Psychopathologie der erworbener Toxoplasmose. Nervenarzt 50: 128-131.

GOLDMAN M. 1957: Staining Toxoplasma gondii with fluorescein-labelled antibody. I. The reaction in smears of peritoneal exudate. J. Exp. Med. 105: 549-556.

HAY J., AITKEN P. P., AMOTT M. A. 1985: The influence of congenital toxoplasma infection on spontaneous running activity of mice. Z. Parasitenkd. 71: 459-462.

HOOGENBOOM I., DIJKSTRA C. 1987: Sarcocystis cernae: a parasite increasing the risk of predation of its intermediate host, Microtus arvalis. Oecologia 74: 86-92.

KAVALIERS M., COLLWELL D. D. 1982: Parasitism, opioid systems and host behaviour. Adv. Neuroimmun. 2: 287-295.

KRAHENBUHL J. L., REMINGTON J. S. 1982: The immunology of Toxoplasma and toxoplasmosis. In: A. J. Mahmias. and J O'Reily (Eds.), Immunology of Human Infections, Part II., Plenum Publishing Corporation, New York, pp. 356-421.

KRAMAŘ J., ČERNÁ Ž. CHALUPSKÝ J. 1963: Immunofluoreszenzreaktionen in der serologischen Diagnostik der Toxoplasmose. Z. Bakter. Parasitol. Infekt. Hyg. 1 Orig. 193: 523-534.

LADEE G. A., SCHOLTEN J. M., POSTHUMUS MEYES F. E. 1966: Diagnostic problems in psychiatry with regards to acquired toxoplasmosis. Psychiat. Neurol. Neurochir. 69: 65-82.

MINTO A., ROBERTS F.J. 1959: The psychiatric complications of toxoplasmosis. The Lancet, June 6: 1180-1182.

MOHAN V., CHOPRA R. 1986: A study of personality variation in women before and after menstruation. Pers. Individ. Diff. 7: 127-128

REMINGTON J.S. 1974: Toxoplasmosis in the adult. Bull. New York Acad. Med. 50: 211-227.

REMINGTON J.S., KRAHENBUHL J.L. 1982: Immunology of Toxoplasma gondii. In: A. J. Nahmias and J. O'Reilly (Eds.): Immunology of Human Infection, Part II., Plenum Publishing Corporation, New York, pp. 327-371.

STEINER M. 1987: The effects of gonadal hormones on brain and behavior. Prog. Neuropsychopharmacol. Biol. Psychiatry 11: 115-119.

Fig. 1.

Effect of toxoplasmosis and age on four personality factors estimated in 195 male subjects by Two-way analysis of variance (BMDP, program P7D). The effects of toxoplasmosis, age and toxoplasmosis-age interactions were tested. Only the tail probabilities for toxoplasmosis effects are given. Open and crosshatched bars show the means with standard errors of a particular personality factor for different age classes of Toxoplasma negative and positive subjects, respectively. The cut points for the age classes correspond to those used for the age standardization of row data obtained by Cattell's questionnaire. The numbers of subjects in groups on the graph were indicated on the upper plot. The equality of variances for the factors listed was confirmed by the Levane's test.