After a bench trial, appellant Alisha Townsend was found guilty of driving under the influence ("DUI"). She argues on appeal that the trial court erroneously admitted testimony regarding the result of a vertical gaze nystagmus ("VGN") test conducted by a Metropolitan Police Department ("MPD") Officer on the day of the incident. We agree that the trial court erred in admitting the testimony. Because we are unable to say with fair assurance that the trial court would have found guilt beyond a reasonable doubt in the absence of the VGN testimony, we vacate the judgment and remand for the court to reconsider its verdict without reference to the VGN test results.
*729I.
At trial, MPD Master Patrol Officer Ronald Carroll testified that on April 2, 2015, he arrived at the 200 block of Missouri Avenue, N.W., in response to a request for assistance made by another MPD officer. Officer Carroll testified that, upon arriving at the scene, he observed appellant "behind the wheel" of a vehicle whose engine was running. The vehicle was partially on the sidewalk and partly in the westbound lane, but facing eastbound. When Officer Carroll approached the driver's side of the vehicle, he asked appellant what she was doing and received a response from appellant that was "incoherent." Officer Carroll testified that appellant "was not really looking at [him]" and, at the same time, "was trying to pull the car off the sidewalk." Officer Carroll attempted to coax "[appellant] to step out of the car," which she refused to do for about "three or four minutes." Once appellant finally exited the vehicle, Officer Carroll asked appellant for her license. Appellant repeatedly responded to this question by saying, "wait, wait, wait, hold up, hold up, hold up. What are you here for?"
Officer Carroll also testified that based on his experience of contact with people under the influence of alcohol or drugs approximately 500 times during the course of his career, his opinion was that appellant "was under the influence the way she was repeating herself and refused to cooperate ... at the time." Officer Carroll additionally told the court that a female officer who had arrived "to help ... do a search of [appellant]" "started checking around the vehicle," and advised Officer Carroll that there was a baby in the vehicle. Officer Carroll then observed the child "in the third seat in the right rear corner" of the vehicle, secured in a baby seat. Officer Carroll stated that "at [no] point during [his] interaction with [appellant], did she mention to [him] her baby's presence in the car." Officer Carroll did not smell PCP, marijuana, or alcohol when he was standing next to appellant.
MPD Officer Marcus Malloy testified that when he arrived on the scene, he encountered appellant, who, he testified, "based on [his] experience[ ]" "was confused" and "appeared ... under the influence of drugs." Officer Malloy explained that appellant "didn't know how she arrived at the location she was at," "[h]er balance just seemed unstable," and "she couldn't answer the questions giving complete answers." Officer Malloy therefore performed standardized field sobriety tests. He told the court that he was "instructed on the N[H]TSA [National Highway Transportation Safety Administration] training which included the horizontal gaze nystagmus test [ ("HGN") ], the walk and turn test, and the one legged stand [test]" and that he was "certified ... to perform the tests." He testified that he began by conducting the walk-and-turn test.1 Appellant "couldn't keep her balance while [Officer Malloy] was giving her instructions," and he "had to place her back into the starting position." On appellant's first attempt, "she took too many steps and ... didn't turn around" or count out loud.
*730When the officer asked appellant "to stop and turn around and take nine steps back," she "missed the heel to toe instructions" and again "did not count out loud." Officer Malloy's observation was that appellant "exhibited eight clues on the walk and turn tests, which mean[t] she approximately had a blood alcohol content of at least a .08."2
In addition to the walk-and-turn test, Officer Malloy administered the one-leg-stand test.3 The results were that appellant "couldn't keep her balance" and "would raise her foot and then ... drop her foot back," "swaying while she performed the test." Officer Malloy testified that he then placed appellant under arrest for driving under the influence.
In explaining why he "came to the conclusion that [appellant] may possibly be under the influence of drugs," Officer Malloy told the court that he had also performed a VGN test on the date of the incident.4 He explained to the court that he conducted the test with appellant's glasses on, even though he knew he was "not supposed to do that" and even though he had learned through his NHTSA training that "if you have glasses on, th[e results] are inaccurate." He demonstrated how he conducted the test, causing the prosecutor to state that the officer held his pen light "just shy of 15 inches" away from appellant's eyes rather than the "8 inches to 12 inches from the eye" that the officer "believe[d]" NHTSA prescribed. Officer Malloy testified that "if [appellant] demonstrated the nystagmus vertically, then more than likely, she's under the influence of some type of narcotic or drug" and told the court that appellant "had an irregular nystagmus," meaning that "the eye ha[d] an involuntary jerking." The court asked questions during Officer Malloy's demonstration and asked the officer "exactly what vertical gaze nystagmus is indicative of." Officer Malloy replied that it is indicative of "[n]arcotics, like PCP."5
At the conclusion of the government's case, appellant's trial counsel moved to strike Officer Malloy's testimony and also for a judgment of acquittal, arguing that the government needed an expert in this case but had not "qualified [Officer Malloy] as an expert" and that the officer had acknowledged doing the VGN inaccurately. The court denied the motions, reasoning that "the only circumstance [where] the [g]overnment needs to b[r]ing a qualified expert in here to testify with respect to the standardized field tests is when the [g]overnment is putting forth evidence of horizontal gaze nystagmus." Reciting the evidence elicited during the government's case, the court concluded that there was "evidence such that a reasonable fact finder at this juncture could find guilt beyond a reasonable doubt."6
In the defense case, appellant testified that her friend (who Officer Carroll testified *731"was standing on the street in front of the vehicle" when the officer arrived on the scene), who had been driving the vehicle that day because appellant "didn't really know ... th[e] area," made a wrong turn and "hit [a] mailbox." Appellant explained at trial that after her friend hit the mailbox, both appellant and her friend exited the vehicle "to go look at the damage," and appellant returned to the passenger side of the vehicle, where she was when police arrived at the scene. Appellant told the court that when Officer Carroll arrived and instructed her to exit the vehicle, she was "telling him to hold on, wait" because she was trying "to get somebody on the phone like [her] dad or somebody to tell [her] what to do from [t]here because [her] daughter [wa]s in the car." Appellant explained that she "was scared out of [her] mind" concerning the "safety of [her] child," and that officers administered the sobriety tests after telling her that she could not go and check on her daughter. Appellant said that at the time when Officer Malloy was instructing her on the field sobriety tests, there was "a lot of confusion" and "there w[ere] several different officers speaking at the same time."
Appellant further testified that in conducting one of the sobriety tests, Officer Malloy asked her to take off her glasses, which she refused to do, telling the officer that she "c[ould]n't see without [her] glasses." Appellant told the court that at the time of the sobriety tests, "her head [wa]s pounding," "[her] back [wa]s hurting" because she has "three herniated disks and two pinched nerves," she "was cramped up" in the car, and she was "scared." Additionally, appellant testified that she "ha[s] arthritis" in her knees and that she had taken her prescribed medications that day, including Percocet, which can "make [her] a little drowsy" and which she took five hours before the incident; Albuterol and Advair, which appellant told the court were "breathing treatments"; and Adderall, which appellant said can make her "talk a little bit more than [she] already ... talk[s]." Appellant testified that she "was a little drowsy that evening," but pointed out that she had "been in the car for 17 hours" (during a road trip to the District of Columbia from Florida).
In announcing its findings and verdict, the trial court began by noting that the government had presented a "totality of the circumstances" theory to prove the DUI charge. The court then incorporated by reference the summary of the evidence it had recounted when ruling on appellant's first MJOA, and thereafter summarized the subsequent testimony. The court noted that under this court's opinion in Karamychev v. District of Columbia , 772 A.2d 806 (D.C. 2001), an expert is required to provide testimony regarding the HGN test, but reasoned that "[t]he vertical gaze and nystagmus test does not require expert qualification." The court also found that the test can be "indicative of high alcohol usage and/or drugs."
The court credited the testimony of Officers Carroll and Malloy.7 Specifically, the court credited Officer Carroll's testimony that appellant "was incoherent" and asked the officer, "what are you here for?" The court likewise credited Officer Malloy's testimony that appellant "appeared confused, incoherent[,] and ... like she was under the influence of something" and "had unstable balance," as well as Malloy's *732testimony regarding what he observed while conducting the standardized field sobriety tests: that during the walk-and-turn test, appellant "t[ook] too many steps, fail[ed] to maintain proper balance, [and] miss[ed] heel to toe"; that during the one-leg-stand test, appellant "couldn't maintain her balance," causing Officer Malloy to discontinue the test; and that during the VGN test, appellant "showed vertical gaze nystagmus present." The court remarked that Officer Malloy's allowing appellant to keep her eyeglasses on during the VGN test "does not invalidate what Officer Malloy observed." The court found that, as the officers had testified, appellant was seated in the driver's seat of the running vehicle, which was in part on the sidewalk facing the wrong direction. The court concluded that appellant's "admission to taking [four prescription] medications along with the observations of the officers ... established beyond a reasonable doubt that [appellant] was under the influence of drugs ... to a degree that was perceived or noticed."
After appellant filed a motion for review by an Associate Judge, the Honorable Zoe Bush affirmed the "ruling and sentence of the Magistrate Court" on March 22, 2016. This appeal followed.
II.
Appellant contends on appeal that the trial court "misapplied Karamychev " and "abused its discretion when it permitted Officer Malloy to testify about vertical gaze nystagmus and the conclusion that its presence is evidence of intoxication by a drug." Appellant argues that allowing "Officer Malloy to testify on this scientific topic and to opine an ultimate conclusion on a central element of the charge ... substantially prejudiced [his] rights, jeopardized the fairness of the trial, and had a potentially substantial impact on the outcome of the trial." The District of Columbia does not defend the trial court's ruling with respect to the VGN test, but argues that "any error in the admission ... [of the] vertical gaze nystagmus test was harmless."
We agree with appellant that Karamychev was not authority for the trial court to admit testimony about the results of the VGN test without a foundation laid by expert testimony. This court's reasoning in Karamychev , that "[t]here is no doubt that the administration of the HGN test and the interpretation of the results are subjects 'beyond the ken' of a lay juror, and expert testimony [i]s therefore required," 772 A.2d at 811-12 (internal citation omitted), applies equally to the VGN test and interpretation of its results. Indeed, one opinion of this court refers to the "vertical nystagmus portion of the HGN test." Sanchez v. District of Columbia , 102 A.3d 1157, 1160 (D.C. 2014). As courts have recognized, and as appellant asserts, both the HGN and VGN test involve the use of "scientific procedures and techniques."8 Both "differ fundamentally from other field sobriety tests."9 This *733does not necessarily mean that evidence about a defendant's results on the VGN test may never be presented without detailed testimony about the scientific theory underlying the test. Cf. Karamychev , 772 A.2d at 812 (citing with approval the holding of several courts that "testimony by a properly trained police officer with respect to the administration and results of the horizontal gaze nystagmatus evidence is admissible without need for further scientific evidence" (internal quotation marks omitted) ).10 It does mean that to testify about the results, the witness must first be shown to have specialized training in administering the test and proficiency in observing its results. Before the VGN testimony could be admitted into evidence, it was necessary that the witness presenting that evidence be offered to the court, and accepted by the court, as an expert in administering the test.
Here, the District of Columbia did not proffer Officer Malloy as an expert with respect to the VGN test and the court did not qualify him as an expert. Further, although Officer Malloy testified about his NHTSA training on the HGN tests and his NHTSA certification to administer standardized field sobriety tests, and although *734he made some references during his testimony to what he had learned about the VGN test, he did not testify specifically that he had completed training on administration of the VGN test or that he had been certified to administer it. He also gave less-than-definite answers to cross-examination questions about what defense counsel suggested were mandated VGN test protocols (referring, for example, to the "8 inches to 12 inches from the eye" that the officer "believe[d]" NHTSA prescribed regarding where the penlight should be held in relation to an examinee's eyes). In short, not only was Officer Malloy not offered or accepted by the trial court as an expert in administration of the VGN test, it is not even clear from the limited foundation that was laid that he could have been qualified as an expert with respect to that test. For all these reasons, we conclude that the trial court erred in admitting the VGN testimony.
The remaining question is whether admission of the VGN testimony, though error, was harmless. The issue is not whether the evidence was sufficient for conviction even without the erroneously admitted evidence-we have no trouble concluding that it was-but rather whether we can "say 'with fair assurance, after pondering all that happened, without stripping away the erroneous action from the whole, that the judgment was not substantially swayed by the error.' " Andrews v. United States , 922 A.2d 449, 458 (D.C. 2007) (quoting Kotteakos v. United States , 328 U.S. 750, 765, 66 S.Ct. 1239, 90 L.Ed. 1557 (1946) ). " 'To conclude that an error is harmless under Kotteakos , we must find it highly probable that that error did not contribute to the verdict.' " Id. (internal brackets omitted) (quoting Wilson-Bey v. United States , 903 A.2d 818, 844 (D.C. 2006) (en banc) ).
The government had the burden at trial of showing beyond a reasonable doubt that appellant "operate[d] or [was] in physical control of [the] vehicle" (a requirement that appellant does not dispute was satisfied by the credited testimony that she was behind the wheel of the vehicle with the engine running) while "under the influence of alcohol or any drug or any combination thereof." D.C. Code § 50-2206.11 (2012 Repl.). A person is guilty of driving under the influence if she is "appreciably"-meaning "capable of being perceived by the naked senses"-"less able, either mentally or physically or both, to exercise the clear judgment and steady hand necessary to handle ... [an] automobile with safety to h[er]self and the public." Taylor v. District of Columbia , 49 A.3d 1259, 1267 (D.C. 2012) (internal quotation marks omitted). To infer impairment, the court may consider "the testimony of officers and other witnesses about their observations of the defendant's appearance, speech[,] and actions." Koonce v. District of Columbia , 111 A.3d 1009, 1017 (D.C. 2015). The government "need [not] prove the particular type of drug that caused appellant's impairment in order to convict h[er] under the DUI statute," Carrington v. District of Columbia , 77 A.3d 999, 1006 (D.C. 2013), and "circumstantial evidence [of impairment by drugs and/or alcohol] will suffice even though it does not specifically quantify the amount of the substance ingested and relate it to the ability to drive." Harris v. District of Columbia , 601 A.2d 21, 26 (D.C. 1991).
In this case, the trial court credited the officers' testimony about numerous factors that supported a conclusion that appellant was under the influence. The vehicle appellant was trying to maneuver was partly on the sidewalk, facing the opposite direction from the flow of traffic. Appellant appeared to the experienced officers to be "confused," like she was "under the influence *735of drugs"; she was incoherent, repeated herself, and did not seem to understand what was going on around her; she was swaying and could not keep her balance during the walk-and-turn and one-leg-stand tests; she "took too many steps," "didn't turn around," and failed to count out loud, and otherwise failed to follow instructions; and, to Officer Malloy's recollection, she offered no medical explanation for her inability to perform the tests as instructed. Further, in testimony that the trial court credited, appellant acknowledged that she had taken a prescription medication, Percocet, which she said made her drowsy. The trial court recounted this evidence at great length, in statements that covered over six transcript pages of the court's ruling on the MJOA.
By comparison, the court devoted only a few sentences of its MJOA ruling, and a sentence of its findings in support of the verdict, to Officer Malloy's testimony about the VGN test results. The court acknowledged the testimony about how the VGN protocol called for the examinee to remove her eyeglasses, which appellant did not do. In addition, the court seemed to recognize that its inquiry was about whether the "totality of the circumstances" proved the DUI charge beyond a reasonable doubt. Further, the prosecutor mentioned the VGN results only briefly in closing argument as one among many factors that led the officers to believe that appellant was under the influence of drugs.
At the same time, the record makes clear that the trial court was interested in the VGN testimony. The court followed up with questions of its own to Officer Malloy about what the VGN test is indicative of. In its findings, the court was satisfied that VGN is "indicative of high alcohol usage and/or drugs." The court specifically noted in its findings that during the VGN test, appellant "showed vertical gaze nystagmus" and further remarked that Officer Malloy's allowing appellant to keep her eyeglasses on during the test "does not invalidate what Officer Malloy observed." And although appellant admitted to having taken Percocet, the strength of any inference that she was Percocet-impaired was perhaps at least somewhat diminished by her testimony that she took the drug five hours before the incident and that she had been on the road for seventeen hours (testimony that the trial court did not specifically discredit).
In light of all these points, we deem it a close question whether the court would have reached the same verdict even in the absence of the VGN testimony. Ultimately, we are persuaded by the likelihood that testimony about the results of a VGN test, like testimony about the results of an HGN test, "carries with it an 'aura of certainty' [such that the finder of fact] may therefore overestimate its probative value." State v. Blackwell , 408 Md. 677, 971 A.2d 296, 308 (2009) (explaining why the court was unable to conclude that erroneously admitted HGN testimony in no way influenced the jury's DUI guilty verdict). We conclude that the proper course is to "remand this case for the court to reweigh the evidence in the record afresh [without considering the VGN testimony], and render a new verdict." Hawkins v. United States , 103 A.3d 199, 202 (D.C. 2014) (internal quotation marks omitted).
Wherefore, the matter is remanded for further proceedings consistent with this opinion.
So ordered .