GREENBERG, J.

From 1983, until his arrest in October, 1992, the defendant lived at home with his wife and family; two children, Ashley and John, were ten and five years old at the time of trial. In October, 1992, Ashley told an investigator from the Department of Social Services that both parents had sexually abused her and her younger brother over a period of years. As a result of these accusations and subsequent police investigation, the children were placed in foster care. Both parents were charged, the defendant on thirty-five indictments for rape of a child (three with force), six counts of indecent assault and battery on a child under fourteen, and two counts of posing or exhibiting a child in a live performance involving sexual conduct. After trial and three days of deliberation, the jury convicted the defendant on eleven indictments for rape of a child and all of the indecent assault and battery and unlawful exhibition of a child charges. The defendant appeals, bringing up for review numerous evidentiary points and other issues.

We conclude that the judge improperly admitted expert testimony on the characteristic profile of an incestuous family and other testimony directly linking the characteristics of sexually abused children to the complainants in this case. Consequently, we reverse the convictions.

The issue that requires our scrutiny pertains to testimony offered by the government's expert, Dr. Denise Gelinas, a psychologist, and Maureen Brown, a school counselor and social worker, regarding "child disclosure processes" and on the behavior patterns of children who have suffered sexual abuse. It is settled that the trial judge has discretion to allow such expert testimony where it may assist the jury in deciding a contested issue, including "expert testimony on the typical symptoms of sexually abused children because the information is beyond the common knowledge of jurors and of assistance in assessing a victim witness's testimony and credibility." Commonwealth v. Dockham, 405 Mass. 618, 629-630 (1989), and cases cited. Though the Dockham court held that testimony about the behavioral signs and symptoms sexually abused children frequently exhibit may be admitted at the judge's discretion, "the line between proper testimony as to patterns of disclosure of child sexual abuse victims and improper testimony constituting endorsement of the credibility of a victim-witness is indeed a narrow one." Commonwealth v. Rather, 37 Mass. App. Ct. 140, 148 (1994). Such limits as bear on the discretion of a trial judge to bar impermissible opinion testimony in child sexual abuse cases were recently discussed in Commonwealth v. Richardson, 423 Mass. 180, 186 (1996). The test of admissibility of sexual profile testimony is whether "the witness explicitly links the opinion to the experience of the witness child." Ibid. Such testimony clearly amounts to impermissible vouching. "Where there is no link between the expert testimony and the victim, the opinion is usually allowed." Ibid., and cases cited. However, in some cases, the Richardson court said that "opinion testimony that does not explicitly link the opinion to the child witness [may] nevertheless constitute impermissible vouching." Ibid., and cases cited.

The case at bar falls into the last category. The line was crossed when, over the defendant's objection, Dr. Gelinas went beyond the description of general principles of social or behavioral science that might assist the jury in their deliberations concerning credibility and gave testimony concerning family dynamics that evolved into profile testimony that signaled the jury that the child complainants were sexually abused. This testimony included her statement that "the most characteristic constellation or structure of an incestuous family is one in which the mother is usually . . . overburdened or impaired in some way. . . . [The father] usually ha[s] in some ways a more active relationship with the children in the family than [the] mother does . . . and . . . that puts the oldest [child], usually daughter, at risk to be sexually abused by the father in a situation in which he feels pretty highly entitled. . . . I almost hate to go into it but, basically, it puts the incestuously abused child in a terrible position because the person she is closest to . . . and the one to whom she probably feels the most loyal is the one who's doing the sexual abus[e]." In essence, the expert gave characteristic sexual profile testimony that presented the defendant's family situation as prone to sexual abuse, suggesting to the jury that this was a reliable factor as to whether sexual abuse occurred. This inadmissible profile evidence had the effect of identifying the defendant as a person likely to commit incestuous sexual abuse, and the complainants as children who testified truthfully to the occurrence of sexual abuse. See Commonwealth v. Ianello, 401 Mass. 197, 201-202 (1987).

Maureen Brown, Ashley's school counselor, over objection, offered her personal observations gleaned from regular counseling sessions with Ashley over several years. Before allowing Brown to give testimony concerning Ashley's behavioral symptoms, the judge said that he was "a little leery of [Brown] giving an opinion on the ultimate issue in this case." However, not having made a definitive ruling at that time, the judge permitted her to testify that "the biggest . . . sign that [she] observe[d] in [Ashley] that [she] could categorize as consistent with [her] experience dealing with traumatized children . . . would be [Ashley's] need to present kind of a blank slate. Very often children who have been raised in abusive homes are clearly given a rule, don't talk, don't trust, don't feel." Over the defendant's objection, Brown identified as "clinical symptoms" Ashley's "[in]ability to deny or repress her feelings," social isolation from her peers, resistance to "therapeutic relationships," "stares or spaceyness," and other diagnostic behaviors. She concluded that "[t]hese . . . things . . . in my experience and in the literature are significant in determining an abused child." Although she did not directly render an opinion as to whether Ashley was sexually abused, we conclude that Brown's testimony amounted to a lay opinion that Ashley was an abused child. Contrast Commonwealth v. Allen, 40 Mass. App. Ct. 458, 467-468 (1996). Her testimony went beyond her classroom observations of Ashley by improperly relating those observations to her experience with other traumatized children. Such testimony borders on a statement as to how this witness feels about the credibility of the child complainant. See Commonwealth v. Trowbridge, 419 Mass. 750, 759 (1995) (doctor's opinion testimony comparing child complainant to general characteristics of sexually abused children impermissibly intruded upon jury's province to assess credibility of the child witness).

Together with Dr. Gelinas's testimony that some children disassociate in order to adapt to sexual abuse and that one of the behaviors observed is that "the child will often look kind of spacey." Brown's testimony acquired the veneer of an expert. The danger of such implicit vouching is greater where, as here, the witness is testifying to actual interaction with the child.

Although Dr. Gelinas was not a treating therapist, somewhat lessening the risk that her testimony could "be construed as impliedly supporting the truthfulness of [the complainants]," Commonwealth v. McCaffrey, 36 Mass. App. Ct. 583, 591 (1994), that was not the case with Brown. Brown's direct testimony about Ashley's demeanor at school and behavior in regular small group discussions was highly suggestive. Coupled with Dr. Gelinas's testimony of behavioral symptoms of sexually abused children, there was a substantial risk that the jury's confidence in Ashley's credibility would be enhanced. This was especially true when the testimony came from her school counselor. See Commonwealth v. O'Brien, 35 Mass. App. Ct. 827, 832 (1994) (warning prosecutors to avoid questions to witnesses of whether child victim's behavior conformed to the "general characteristics" of abused children). See also Commonwealth v. McCaffrey, supra at 592 (treating therapist who testifies to general syndromes might be viewed as implicitly endorsing the complainant's testimony by virtue of the fact that she has accepted complainant into therapy).

In a recent decision, Commonwealth v. Brouillard, 40 Mass. App. Ct. 448, 453 (1996), we reviewed the testimony of a treating therapist who exceeded the permissible testimonial boundaries by explicitly connecting the complainants to the general syndromes associated with sexual abuse. In Brouillard, we also concluded that the evidence "had the effect of impermissibly vouching for the complainants' credibility." Id. at 451. As in Brouillard, the combination of Gelinas's expert testimony and Brown's testimony, while not directly vouching for the complainants' credibility nor expressly connecting the children's claim of sexual abuse to the defendant, served as the functional equivalent. See Commonwealth v. Trowbridge, 419 Mass. at 759-760. We conclude that such testimony is inadmissible because it usurps the jury's traditional functions. Commonwealth v. Ianello, 401 Mass. at 202.

Having concluded that the direct testimony on credibility should not have been admitted, we turn now to the question of whether there was reversible error. Commonwealth v. Montanino, 409 Mass. 500, 504-505 (1991). Part of that determination turns on what the prosecutor argued to the jury. We reprint in the margin a portion of the closing argument in which she urges the jury to consider whether the defendant's wife had a motive to lie. The argument pushed nearly all of the "hot-button" issues and matched the inadmissible expert testimony. We conclude that the prosecutor's closing statements regarding Dr. Gelinas's testimony compounded the problem. The prosecutor's remarks suggested to the jury that the defendant sexually abused his children simply because he fit the mold, e.g., that he was married to a retarded woman, placing him in the position of being the more responsible, more powerful parent. See Commonwealth v. Day, 409 Mass. 719, 725, 726 n.6 (1991) (admission in evidence of child battering profile is reversible error since a reasonable jury could have inferred that the prosecutor was implying that the defendant fit the "child battering profile," and, as a result was responsible for the child's fatal injuries).

To make matters worse, in another part of her argument the prosecutor emphasized that neither Dr. Gelinas nor the counsellor, Brown, had a motive to lie. There was no suggestion by the prosecutor that the witnesses' opinion of the credibility of the complainants would ordinarily be inadmissible, and the judge's charge merely repeated "boilerplate" instructions concerning credibility and the nature of expert testimony. Without more, there was a substantial likelihood that the jury would consider the sexual profile and opinion testimony as supportive of the accuracy, reliability, and truthfulness of the children's testimony. Commonwealth v. Rather, 37 Mass. App. Ct. at 150-151.

Guilt or innocence on these charges ultimately turned on the question of the children's credibility; evidence of sexual abuse was not overwhelming, and physical evidence of abuse was notably minimal. It cannot be said that the jury would have convicted even in the absence of the error in admitting the profile testimony and testimony linking characteristics of the child complainants to characteristics of sexually abused children. Commonwealth v. Rather, supra at 149.

Other issues. The other points raised by the defendant are either without merit or are unlikely to arise at any retrial. With regard to the defendant's argument regarding the denial of his motion for a court-ordered polygraph test, we note that the Supreme Judicial Court has given no indication that it intends to abandon the formulation in Frye v. United States, 293 F. 1013 (D.C. Cir. 1923), in the polygraph contest. See Commonwealth v. Mendes, 406 Mass. 201, 205 (1989).

Judgments reversed.

Verdicts set aside.