The defendant, Daniel G. Chamberlain, Jr., was convicted of two counts of aggravated felonious sexual assault, RSA 632-A:2 (1986 & Supp. 1990) (amended 1992), after a jury trial in the Superior Court (Mohl, J.). We reverse one of the convictions because it was not sufficiently supported by the evidence presented at trial. We reverse the remaining conviction in light of our opinion in State v. Cressey, 137 N.H. 402, 628 A.2d 696 (1993), on the ground that the testimony of the State’s expert witness was improperly admitted at trial.
The defendant was indicted in 1990 on two counts of aggravated felonious sexual assault against his niece. The first indictment alleged that the defendant digitally penetrated the child’s vagina, and the second indictment alleged that he engaged in cunnilingus. The charges in the first indictment arose from an incident during which the defendant was rubbing the child’s vagina while purportedly applying ointment to a rash. The only evidence offered by the State to establish digital penetration was the testimony of the child victim. Her testimony proceeded as follows:
*416“Q. Where was the rash?
A. On my vagina.
Q. Okay. Did Mr. Chamberlain put something on the rash?
A. Yeah, lotion.
Q. And how did he put it on?
A. With his finger.
Q. And where did he put his finger?
A. Not inside of me, just on the vagina.
Q. On the outside of your vagina?
A. (Nodding).
[Objection]
Q. .... You’re familiar with your female anatomy?
A. Yes.
Q. When you said he didn’t put it inside of you, what did you mean?
A. He didn’t put it in the opening.
Q. Where did he put it?
A. On the sides.
Q. On the sides of your opening?
A. Yeah.”
The defendant argues that this testimony fails to establish that penetration occurred. We agree.
Sexual penetration is a material element of any aggravated felonious sexual assault offense under RSA 632-A:2. Sexual penetration is defined to include “[a]ny intrusion, however slight, of any part of the actor’s body or any object manipulated by the actor into genital or anal openings of the victim’s body.” RSA 632-A:l, V(e) (1986). In assessing the sufficiency of the evidence presented with respect to penetration, we will view the evidence and all reasonable inferences drawn therefrom in a light most favorable to the State. State v. Baker, 135 N.H. 447, 449, 606 A.2d 309, 310 (1992). We will uphold the conviction on the evidence presented if “a reasonable jury could have found guilt beyond a reasonable doubt.” State v. O’Neill, 134 N.H. 182, 185, 589 A.2d 999, 1002 (1991) (quotation omitted).
We hold that no reasonable jury could have found the defendant guilty beyond a reasonable doubt of the charge alleging digital *417penetration. The evidence presented to establish penetration failed to do so. The child victim not only failed to testify that the defendant had penetrated her, she explicitly stated that the defendant did not put his finger in her opening. The State contends that the child’s representation that the defendant put his finger on the sides of her opening is sufficient to establish penetration beyond a reasonable doubt. Even if the child’s representation could be interpreted so broadly as to imply that the defendant touched the internal sides of her genital opening, this mere implication, in light of the child’s entire testimony, does not permit a rational finding of guilt beyond a reasonable doubt. Therefore, we reverse the defendant’s conviction on the charge alleging digital penetration.
We now turn to the issue of whether the testimony of the State’s expert witness was properly admitted at trial. During its case-in-chief, the State presented the testimony of Ramona Belanger. The child victim in this case had been referred to Belanger for counsel-ling. Belanger testified about her work in the area of child abuse, discussed the symptoms and behaviors she found in the child victim, and ultimately concluded that the child’s symptoms were consistent with those of a child who had been sexually abused.
We addressed the questions surrounding the use of expert testimony in a child sexual abuse prosecution in State v. Cressey, 137 N.H. 402, 628 A.2d 696 (1993). In Cressey we held that the expert testimony of a psychologist, based on her psychological evaluations of the child victims, was not reliable enough to be offered as evidence that the children had indeed been sexually abused. Id. at 407, 628 A.2d at 699; cf. Daubert v. Merrell Dow Pharmaceuticals, Inc., — U.S. —, 113 S. Ct. 2786 (1993). We noted, however, that certain expert testimony could be offered to preempt or rebut an inference, based on the child victim’s actions and behaviors following the alleged abuse, that the child has lied about the abuse. Cressey, 137 N.H. at 412, 628 A.2d at 703. In instances of sexual abuse, child victims may commonly exhibit certain behaviors related to the disclosing of abuse that are potentially subject to misinterpretation by a lay jury. See State v. J.Q., 252 N.J. Super. 11, 28-31, 599 A.2d 172, 181-83 (1991), aff’d, 130 N.J. 554, 617 A.2d 1196 (1993); State v. Spigarolo, 210 Conn. 359, 377-78, 556 A.2d 112, 122-23, cert. denied, 493 U.S. 933 (1989). Such commonly observed behaviors include the delayed reporting of abuse, inconsistent recountings of the abuse, and recantation of the initial disclosure. See Cressey, 137 N.H. at 411, 628 A.2d at 702; Spigarolo, 210 Conn. at 377, 556 A.2d at 122. These behaviors often may be the result of the trauma experienced by the child and of *418the unique dynamic existing in child sexual abuse cases. See State v. J.Q., 252 N.J. Super. at 28-29, 599 A.2d at 181. For example, a child who is sexually abused by a close relative, or has been threatened into silence by the abuser, may be reluctant to disclose the abuse and unsure about pursuing a complaint. Id. Without an understanding of the reasons behind these behaviors, a jury may automatically infer from a child’s secrecy, inconsistency, or recantation that the child has fabricated his or her testimony. Therefore, when a child’s actions after an alleged incident of sexual abuse have the potential to lead a jury to conclude that the child is lying, the testimony of a qualified expert may be beneficial to offer an alternative explanation for the child’s specific behavior so that the jury may more accurately judge the credibility of the child victim. Cressey, 137 N.H. at 411-12, 628 A.2d at 702-03.
In the present case, the testimony of Ramona Belanger was largely based on the child sexual abuse accommodation syndrome. She identified several characteristics, such as secrecy, helplessness, accommodation, and incomplete disclosure, that are part of the syndrome. The purpose of her testimony, however, was not to educate the jury about the characteristics and offer an explanation for some of the child victim’s behaviors. The purpose of her testimony was to prove that the child victim had been abused by showing that she exhibited behaviors and characteristics identical to those identified by the child sexual abuse accommodation syndrome. Belanger concluded her testimony for the State by confirming that the behaviors of the child victim were consistent with those of a child who had been sexually abused.
Our holding in Cressey specifically prohibits an expert from testifying for this purpose in a criminal child sexual abuse prosecution. Cressey, 137 N.H. at 407, 412, 628 A.2d at 699-700, 703. Moreover, even if an expert could so testify, we recognized in Cressey that the child sexual abuse accommodation syndrome cannot properly be used as a diagnostic device to detect whether a child has been sexually abused. Id. at 411, 628 A.2d at 702 (citing State v. J.Q., 252 N.J. Super. at 28, 599 A.2d at 181). The expert testimony we considered admissible under Cressey may be offered only to preempt or rebut an inference from some specific behavior, action, or inconsistency displayed by the child victim that otherwise may be misinterpreted by a jury. Id. at 412, 628 A.2d at 703. The child victim in the present case, for example, made several inconsistent statements during the course of the State’s investigation and, at one point, threatened to retract her accusations. To the extent these behaviors are commonly exhibited by *419sexually abused children, qualified expert testimony, if otherwise admissible, could have been elicited to offer an alternative explanation for these behaviors to rebut the potential inference that she had fabricated her entire account of the abuse.
The expert testimony offered by the State in this case went well beyond what we have allowed in Cressey. After reviewing the record, we cannot say beyond a reasonable doubt that the expert testimony did not affect the verdict, see State v. Elwell, 132 N.H. 599, 607, 567 A.2d 1002, 1007 (1989), and, therefore, we do not find the error harmless. We reverse the defendant’s conviction on the charge alleging cunnilingus and remand for a new trial on that charge. In light of our holdings in this case, we do not address the other issues raised by the defendant in his brief.
No. 91-S-080 reversed; no. 91-S-078 reversed and remanded.
All concurred.