Past impressions: prior relationships cast a long shadow over our social lives.

In a 1948 book, psychoanalyst Theodore Reik described an extraordinary "Dr. Jekyll and Mr. Hyde"-type identity change that he underwent in the minds of many patients during therapy sessions. At the start of each encounter, Reik wrote, patients perceived a bald, elderly man with a big nose and glasses who presented a thoughtful, friendly demeanor. In other words, they saw the therapist for who he actually was.

In the heat of therapy sessions, however, the real Reik disappeared from the patient's mind. Referring to himself, the clinician wrote in Listening with the Third Ear (Farrar, Straus, and Co.) that "during the past hour the patient may have been considering this same man as near to God or close to Satan; he may have seen in him his grandfather or father or a representative of any one of the figures that played an important role in his life."

That's a heavy burden to carry, but psychotherapists of all theoretical stripes bear it regularly. About a century ago, Sigmund Freud, the father of psychoanalysis, dubbed this phenomenon transference. He portrayed transference as a process in which patients unconsciously overlay past relationships onto current ones. Most commonly, Freud theorized, an individual will shift childhood fantasies and sexual conflicts with parents onto his or her analyst.

Some psychoanalysts after Freud have viewed transference as applying not only to therapists. They see it in the reenactment with new individuals of patterns that were established with key people in one's life. The phenomenon is grounded in a need to regularly forge satisfying and secure social ties, those analysts say.

Until recently, however, transference remained unexamined by researchers. Psychoanalytically oriented therapists shunned science as too crude to illuminate the complex inner workings of the mind. Scientists dismissed transference as a fuzzy, Freudian conjecture. Meanwhile, in the clinical realm, the growing popularity of cognitive-behavioral psychotherapy put a premium on dealing with problems in daily life rather than on exploring the relationship between patient and therapist, further marginalizing transference.

Now, the l00-year-old concept is showing signs of renewal. This revival springs from laboratory research in which experimenters trigger transference responses in college students. In such work, subtle reminders of key relationships from the past influence volunteers' self-images and their first impressions of others. Investigators are beginning to unravel the emotional fallout of negative-transference reactions, such as those displayed by women who endured childhood abuse.

"Transference may be ubiquitous in people's everyday interpersonal interactions and important relationships, says psychologist Susan Andersen of New York University. "It can lead to emotionally painful consequences or to feeling connected, bonded, and comfortable:

This research has spurred mental-health clinicians to examine how transference connects psychoanalysis to brain science. Investigations might even explain how numerous forms of psychotherapy work. "You don't have to be psychoanalytic to make use of transference in psychotherapy" says psychologist Drew Western of Emory University in Atlanta.

In addition, a landmark clinical study surprisingly indicates that psychotherapy aimed at confronting transference issues especially benefits severely disturbed persons, who have typically been regarded as poor treatment prospects.

MEMORY TRANSFERS Over the past 15 years, Andersen has chased transference out of clinical hiding and onto the scientific stage. Her pioneering

research pivots on a widespread assumption among mind scientists: Feelings and interaction patterns associated with significant people in one's life can be quickly cued up in new situations. As a result, people learn to view others, and new acquaintances in particular, through a lens of accumulated knowledge about crucial figures from the past.

Andersen's studies employ a two-session, transference-generating technique. In the first session, participants choose two important people in their lives and describe positive and negative characteristics of those individuals, usually by completing sentences provided by experimenters.

Two weeks later, volunteers go to a different lab for an allegedly separate study. An experimenter tells them that they will meet a new person for a "getting-acquainted" conversation. The volunteers then read a series of descriptions of the other in dividual, which they take to be based on a researcher's interview with that person, and are asked to imagine how the upcoming encounter will unfold.

In a session designed to spark transference, participants receive descriptions in which roughly half of the statements paraphrase characteristics that they previously attributed to a key person in their own lives. In a comparison session, each volunteer receives a description peppered with traits of a stranger.

In both cases, descriptions of the new person contain both positive and negative attributes, regardless of how much the volunteer may have loved or loathed the individual on whom the description is based.

In the transference situation, a typical participant endows the unseen stranger with many of the characteristics originally ascribed to the key person in his or her life--regardless of whether those characteristics are in the researcher's description of the supposed stranger.

For instance, a volunteer may instantly form a liking for, or feel safe in meeting, a new person who calls to mind a beloved person from the past. In contrast, the same participant may immediately dislike or feel threatened by a new person with qualities of a hated person from the past.

No such reactions occur for individuals who haven't been coaxed into experiencing transference.

Andersen and her coworkers find that volunteers who experience transference reactions initially display fleeting facial expressions that reflect either positive or negative feelings about the person who has been called to mind. For instance, someone who meets the object of a positive transference often flash a brief smile. Psychotherapists could identify transference in their patients by noting such facial responses, the researchers suggest.

In a 2000 study, Andersen and a colleague reported that transference responses not only affect the individual experiencing it but also alter the behavior of the person who is the target of the transference. In the trial, participant A would briefly speak on the telephone with randomly chosen and unfamiliar participant B, whom researchers had described as similar to an important person in A's life. If, during the conversation, A reacted to B in a manner indicating negative transference, then B would become unpleasant and antagonistic toward A. Pleasant banter dominated during positive-transference encounters. In neither ease did B have any idea that transference had occurred.

"Something in the nonverbal behavior of the speaker [experiencing transference] may influence the partner, such as pauses in speech, a bit of a monotone, or perhaps a lack of enthusiasm; Andersen posits.

The New York psychologist's latest research explores negative transference among women who reported childhood physical and psychological abuse--but rarely sexual abuse--by a parent. Participants expected to meet someone who either did or didn't possess characteristics of the abusive parent and were then told that this person was becoming increasingly tense and irritable.

In the transference condition, women reported disliking and mistrusting the new person. Yet they also cited a sense of indifference and a decline in unpleasant feelings after learning of the new person's irritation, perhaps reflecting the emotional numbing that occurs among abuse survivors, Andersen suggests. The women showed no such responses toward people who did not spark transference.

Overall, the findings reflect an internal tug-of-war between women's feelings of love and hate for abusive parents, Andersen says. For instance, upon hearing about a new person who called to mind an abusive parent, women briefly displayed positive facial expressions before their negative attitudes and feelings bubbled to the surface.

Negative transference apparently assumes many forms. In another study accepted for publication, Andersen studied volunteers who felt either that they had not lived up to a parent's hopes and dreams for them or that they had not fulfilled duties and obligations to a parent. Members of the first group became increasingly sad and distraught as they prepared to meet someone who resembled the parent, apparently because this situation reminded them of a depressing discrepancy between parental aspirations and actual accomplishments. Those in the second group felt increasingly tense and resentful as they waited to meet someone who resembled the parent, since confronting a discrepancy between one's actual behavior and what one ought to have done typically provokes anxiety.

Andersen is now collaborating with psychiatrist Andrew J. Gerber of Columbia University to investigate patterns of brain activation that occur when a person meets someone who resembles an important figure from the volunteer's past.

OLD AND NEW The varieties of transference cultivated in Andersen's experiments mirror how transference works in psychotherapy, according to Emory's Westen. "Patients do not have a transference [to the therapist]; he says. "They have many transferences over the course of a treatment."

In a 2002 paper, Westen and psychiatrist Glen O. Gabbard of Baylor University College of Medicine in Houston described how developments in brain and mind science support this position. A variety of evidence now suggests that the brain continually maps current experiences, such as interactions with new people, onto prior ones--namely, the thoughts, feelings, motivations, and relationship styles associated with important people from the past.

Transference isn't a process in which old social experiences get transferred wholesale into new ones, Westen and Gabbard propose. Instead, it integrates familiar ways of relating to others into current relationship patterns, providing a flexible framework for navigating the social world.

Consider one of Gabbard's former patients, whom he refers to as Ms. C. This 30-year-old, single woman worked as a high-level administrator in a government agency. She attended four sessions per week of psychoanalytic psychotherapy because she felt deeply conflicted about succeeding at work and about becoming romantically intimate with men.

Ms. C's transference reactions to Gabbard took two forms. When the therapist made comments about her relationship with him--such as noting that she seemed afraid of being criticized if she spoke her mind--it triggered transference responses related to her mother. Ms. C's mother had regularly criticized and humiliated her when she expressed her hopes and dreams. To share her inner thoughts with anyone else risked incurring similar reprisals, so she emotionally shut down when the therapist encouraged introspection, and she accused him of being self-centered.

Ms. C's relationship with her father, a generally passive and indifferent figure, lay behind another type of transference. When Gabbard brought up the woman's tendency to overspend and go into debt, she reacted with amusement and questioned his motives in asking such a question. In the past, money mismanagement had been one of the few issues that would elicit angry lectures from her father, a detail-oriented accountant. The father's hostile reactions had nevertheless reassured the woman that he cared about her.

Westen and his colleagues have now measured key aspects of transference exhibited by psychotherapy patients. In a 2005 investigation, 181 psychologists and psychiatrists completed questionnaires about their relationships with randomly selected patients to whom they provided psychotherapy.

Participants described five basic ways in which patients related to psychotherapists: by making excessive demands of the therapist while being angry and dismissive; by fearing the therapist's rejection and compliantly waiting for directives; by talking openly and fostering a good working relationship; by avoiding any feelings for or dependence on the therapist; and by acting seductively toward or feeling sexually attracted to the therapist.

Responses from the same psychotherapists also provided insights into their basic ways of reacting to patients, a process known as countertransference. The therapists' responses consisted of feeling overwhelmed, helpless, positive, overinvolved, sexually attracted, disengaged, protective, and mistreated.

"Transference phenomena are neither mysterious nor unmeasurable" Westen says.

SURPRISING INSIGHTS Psychotherapists have argued for decades about whether treatment should delve into the transference reactions that the treatment itself provokes. Psychoanalytically oriented clinicians believe that a focus on conflicts and themes in a patient's relationship to the therapist illuminates a broad range of personal problems. Another perspective, especially among professionals who provide brief psychotherapy, holds that talk about transference makes patients overly anxious, especially if they're emotionally unstable to begin with.

A new study, directed by psychiatrist Per Hog, lend of the University of Oslo finds, surprisingly, that patients who relate poorly to others and display severe emotional problems--usually considered the poorest prospects for psychotherapy--can be aided by discussing their own transference tendencies.

In their one-of-a-kind study, Hoglend and his coworkers randomly assigned 100 patients to 1 year of weekly, audiotaped psychotherapy sessions in which therapists either did or didn't talk about transference issues. Patients suffered from depression, anxiety, personality disorders, and interpersonal troubles. Therapists generally focused on dealing with interpersonal conflicts and emotional trouble-spots, regardless of whether they mentioned patients' transference reactions.

Before treatment began and after it ended, interviewers assessed the quality of patients' relationships, their emotional health, their capacity for achieving personal insights, and their ability to solve daily problems.

Hoglend's team found that people characterized by unruly emotions and unstable relationships improved to a substantially greater extent if their therapists regularly interpreted transference reactions rather than avoid such discussions. In contrast, patients with mild problems did slightly better in therapy free of transference talk. Results from the Norwegian study appeared in the October 2006American Journal of Psychiatry and in the March 2007 Psychotherapy Research.

Overall, the investigation shows that emotionally fragile patients can profit from hearing someone explain how their past relationships influence their current attitudes and behaviors, remarks Baylor's Gabbard. Transference interpretations may bolster an initially uneasy working relationship between severely disturbed patients and their therapists, thus boosting treatment effects, he speculates.

Hoglend's findings also demonstrate that different facets of psychotherapy promote emotional health in different patients, Gabbard adds. "We should adjust the treatment to the patient, not the patient to the treatment," he says.

Similarly, after a century of clinical confinement, the concept of transference appears finally to have adjusted to the scientific treatment.