In February 1998, Plaintiff Katherine L. Tsoulas ("Tsoulas") became an employee of Medaphis Corporation ("Medaphis") as a Manager of Educational Services. A benefit of her employment was insurance coverage under a group disability policy ("the Plan") provided by Liberty Life Assurance Company of Boston ("Liberty"). Liberty is both the insurer for the Plan and the claims fiduciary.1 Tsoulas applied for and received short-term disability benefits beginning on September 15, 1998 because symptoms of Multiple Sclerosis ("MS") had rendered her unable to work. In March 1999, she applied for and received long-term disability benefits until Liberty discontinued them on March 31, 2004.

Tsoulas filed this action in Maine Superior Court against defendant Liberty, alleging the wrongful termination of her benefits under 29 U.S.C. § 1132, the civil enforcement section of the Employee Retirement Income Security Act ("ERISA"). Liberty removed the action to federal court on November 8, 2004. The United States District Court for the District of Maine granted Liberty's Motion for Judgment on the Stipulated Record on October 27, 2005. Tsoulas now appeals.

We describe the stipulated facts of record. In April 1998, Dr. Kathryn Isaac ("Dr. Isaac"), a staff physician in the Department of Neurosciences at the Stanford University Medical Center, performed the first neurological evaluation of Tsoulas. Dr. Isaac's evaluation revealed that Tsoulas was experiencing "worsening tremors" and blackouts. Other symptoms included "numbness down the right leg and arm" and "fatigue." Dr. Isaac reported that Tsoulas's tremor "appears to be psychogenic" and opined that she would "clearly need psychiatric consultation as well." Magnetic Resonance Imaging (MRI) scans from 1995 and 1997 exhibited two to three brain lesions of unclear etiology. Dr. Isaac reviewed them with a neuro-radiologist, who concluded that the lesions were "not completely typical for multiple sclerosis." Dr. Isaac concluded that she did not believe Tsoulas's symptoms were consistent with a diagnosis of MS.

Dr. Leslie Dorfman ("Dr. Dorfman"), Professor of Neurology and Director of the MS Clinic at the Stanford University Medical Center, examined Tsoulas in May of 1998 and concluded that "[a]lthough the diagnosis is not completely certain, I think it is likely on the basis of the clinical history and the MRI findings that this woman has relatively mild multiple sclerosis, relapsing-remitting form." In addition, Dr. Dorfman determined that Tsoulas was "suffering from an anxiety disorder, and that some of the symptoms which are currently most troublesome to her are related to this, rather than to her MS." Dr. Dorfman recommended follow-up by a neurologist for MS, potential use of corticosteroids for MS exacerbations, and "ongoing concurrent treatment for her anxiety disorder."

In September of 1998, Tsoulas saw Dr. Annette Langer-Gould ("Dr. Langer-Gould"), Acting Director of Stanford University Medical Center's MS Clinic. Dr. Langer-Gould concluded that "although she clearly has multiple sclerosis, she also has a psychiatric condition for which she needs to continue seeking treatment. Initially she refused to do this and . . . vehemently denied that she had any psychological problems." One month later in a follow-up visit, Dr. Langer-Gould wrote that Tsoulas was suffering from MS and a conversion disorder.2 Although Tsoulas was "clearly . . . under tremendous psychological stress," Dr. Langer-Gould told her it was appropriate for her to return to work on a part-time basis. Tsoulas responded that she had discussed the possibility of part-time work with her employer and that it was not an option.

Based on her physicians' recommendations, Tsoulas saw Joyce Brothers Kart, M.F.C.C. ("Ms. Kart") for psychotherapy on four occasions beginning in January 1999 and ending in March 1999. Ms. Kart observed that "[i]t has been extremely difficult for her to function normally . . . . It is very unclear when she will be able to resume active employment due to her symptoms." There is no evidence that Ms. Kart or any other mental health professional has evaluated Tsoulas since March 5, 1999.

In July 1999, Dr. Cathleen Miller ("Dr. Miller") examined Tsoulas and concluded that Tsoulas had "relapsing/remitting multiple sclerosis," but that Tsoulas was "alert and attentive with normal speech and language, and no gross cognitive deficits." Her self-reported symptoms included "episodic worsening of her tremulousness associated with heightened anxiety and bowel incontinence as well as fatigue and impaired balance." Dr. Miller observed that "her current tremor is difficult to characterize, and there is a degree of embellishment noted on examination today." Nevertheless, Dr. Miller wrote to Tsoulas's primary care physician, Dr. Leslie Finder ("Dr. Finder"), that Tsoulas was "greatly disabled at this time of our first meeting."

In October 1999, Liberty sent a questionnaire to Dr. Finder seeking information about Tsoulas's disability. Dr. Finder reported that Tsoulas's primary diagnosis was MS, and that he expected that she would "never" regain either full or modified function. In February 2001, Liberty sent a nurse, Christine Entrekin ("Nurse Entrekin"), to conduct a home visit with Tsoulas to assess her ongoing disability. Nurse Entrekin observed that Tsoulas was "significantly impaired due to physical limitations as well as [Tsoulas's] own assertions of deteriorating mental status. Prognosis for recovery does not appear favorable."

In August 2001, Tsoulas began to receive chiropractic care from Dr. Deborah Baker ("Dr. Baker"). Dr. Baker placed some restrictions on Tsoulas's ability to reach, grasp, pull, push, climb, kneel, bend, squat, walk, stand, conduct repetitive motions, and lift between ten and forty pounds. The only complete restriction she imposed was against lifting forty or more pounds.

In October 2001, Dr. Susan O'Connor ("Dr.O'Connor"), a surgeon, examined Tsoulas in connection with a palpable abnormality of her left breast. Dr. O'Connor removed the abnormality on October 26, 2001, and the diagnosis of breast cancer was confirmed. After the third and final post-operative checkup on August 21, 2002, Dr. O'Connor reported that Tsoulas no longer had any physical restrictions in connection with her breast cancer diagnosis.

As part of its ongoing review of Tsoulas's claim for continued long-term disability benefits, Liberty began planning to conduct surveillance in January 2004. Byers Confidential Investigations ("BCI") conducted the first surveillance on five consecutive days from February 10 through February 14, 2004. On February 12, while the BCI surveillance was ongoing, Tsoulas submitted responses to a routine activities questionnaire indicating the following: 1) she was completely unable to walk or stand without the assistance of a cane, a wheelchair, or a scooter; 2) she spent fourteen to eighteen hours in bed each day; 3) she was able to drive a car only "very little"; 4) she left the house "zero to one" times per week; 5) she never left the house on weekends; 6) she never went shopping at the mall; and 7) she accomplished grocery shopping only with the aid of her son.

The surveillance videos revealed that between February 10 and 14, Tsoulas 1) drove her car unassisted; 2) walked without the aid of a cane, wheelchair, or scooter; 3) completed errands by herself and without any assistive devices; 4) walked down a few steps; 5) conducted banking activities without any assistance; 6) shopped at the mall and the video store; and 7) went grocery shopping, pushed a grocery cart, and loaded and unloaded groceries.

Because there appeared to be significant discrepancies between Tsoulas's self-reported limitations and the video footage, Liberty determined that additional surveillance was necessary. Omega Insurance Services ("Omega") conducted surveillance from March 10 to March 14, 2004. Omega's report revealed Tsoulas: 1) traveling to a tanning salon, a bank, a donut shop, a nail salon, and back to her residence in a single day; 2) walking, driving, entering and exiting a car, and entering and exiting various establishments without the assistance of any braces, supports, or other devices; 3) traveling to a hotel, a parking garage, a restaurant, a comedy club, a night club, and back to the hotel on a single day; 4) traveling to a restaurant, a furniture store, various gift shops, a coffee shop, and a residence in a single day.

After Omega's surveillance was completed, Liberty referred Tsoulas's file to independent consulting physician Dr. John Holbrook ("Dr. Holbrook"), who concluded, based on his examination of her clinical file and the surveillance information, that the diagnosis of MS had "not been conclusively established" and that the majority of Tsoulas's symptoms were "either self-reported, unwitnessed, or unable to be objectively validated." He stated that although she might have undergone a period of psychological stress between 1998 and 1999, recent evidence suggested that she had recovered sufficient functional capacity to be "capable of at least full-time light work."

In March of 2004, Liberty wrote a letter to Tsoulas, terminating her long-term disability benefits, effective March 31. The letter detailed the discrepancies between Tsoulas's self-reported limitations and her actual abilities, as captured by the surveillance videos. The letter also summarized Dr. Holbrook's findings.

On April 30, Tsoulas appealed the termination of her benefits and submitted new material to Liberty to bolster her claim. Liberty then forwarded Tsoulas's file to a second independent consulting physician, Dr. Eric Erlbaum ("Dr. Erlbaum"). Dr. Erlbaum reviewed the file and the surveillance videos, concluding that "the primary diagnosis is most likely psychiatric in nature" and the "second diagnosis is a mild form of multiple sclerosis." Dr. Erlbaum opined further that the surveillance video "would suggest that the claimant is available for full time work of her choosing" and that "her physical condition appears to play a very small role, if any, in limiting her physical capacity." In a letter dated August 17, Liberty denied Tsoulas's appeal of its termination of her long-term disability benefits. Liberty pointed out that although Tsoulas might have significant psychiatric issues, she had not presented a psychiatric record and she was not claiming a disability based on a psychiatric impairment. Without such a record, Liberty was unable to determine whether she had a psychiatric impairment preventing her from meeting her responsibilities "as an Education/Training Manager" at Medaphis.

Tsoulas claims that the district court erred by granting judgment for Liberty on the stipulated record. First, she maintains that the district court applied the wrong standard of review to Liberty's termination of her benefits. Second, she argues that Liberty's decision to terminate her benefits was not supported by substantial evidence. We will address each argument separately.

Our standard of review is governed by the fact that the parties submitted this case to the district court based on a stipulated record. Ordinarily, because summary judgment motions require the trial judge to make legal determinations, we review summary judgment decisions de novo. Reich v. John Alden Life Ins. Co., 126 F.3d 1, 6 (1st Cir.1997). However, we have held that, "[i]n a nonjury case, when the basic dispute between the parties concerns only the factual inferences that one might draw from the more basic facts to which the parties have agreed . . . `[t]he standard for appellate review . . . shifts from de novo review to clear-error review.'" García-Ayala v. Lederle Parenterals, Inc., 212 F.3d 638, 643-44 (1st Cir. 2000) (quoting United Paperworkers Int'l Union, Local 14 v. Int'l Paper Co., 64 F.3d 28, 31 (1st Cir.1995)). The Supreme Court has articulated the clearly erroneous standard thus:

If the district court's account of the evidence is plausible in light of the record viewed in its entirety, the court of appeals may not reverse it even though convinced that had it been sitting as the trier of fact, it would have weighed the evidence differently. Where there are two permissible views of the evidence, the factfinder's choice between them cannot be clearly erroneous.

Anderson v. City of Bessemer City, N.C., 470 U.S. 564, 573-74, 105 S. Ct. 1504, 84 L. Ed. 2d 518 (1985). We review the district court's legal conclusions de novo, Lockheed Martin Corp. v. RFI Supply, Inc., 440 F.3d 549, 552 (1st Cir.2006), but we will set aside the district court's factual inferences "only if they are clearly erroneous." United Paperworkers, 64 F.3d at 31.

When a denial of insurance benefits is challenged under ERISA, courts review the insurer's decision de novo "unless the benefit plan gives the administrator or fiduciary discretionary authority to determine eligibility for benefits." Firestone Tire and Rubber Co. v. Bruch, 489 U.S. 101, 115, 109 S. Ct. 948, 103 L. Ed. 2d 80 (1989). Where, as here, the plan confers such authority upon the fiduciary, the decision is upheld "unless it is `arbitrary, capricious, or an abuse of discretion.'" Wright v. R.R. Donnelley & Sons Co. Group Benefits Plan, 402 F.3d 67, 74 (1st Cir.2005) (quoting Doyle v. Paul Revere Life Ins. Co., 144 F.3d 181, 183 (1st Cir. 1998)).

Tsoulas claims that the district court improperly applied the deferential "arbitrary and capricious" standard to Liberty's termination of her benefits. Specifically, Tsoulas argues that because Liberty is both the insurer and the claims fiduciary — and any finding of eligibility means Liberty will need to pay benefits out of its own pocket — its decision was compromised by an inherent conflict of interest. Citing Wright, Tsoulas alleges that Liberty's decision should have been subject to a heightened standard of review because, "if a court concludes there is an improper motivation amounting to a conflict of interest, the court may cede a diminished degree of deference — or no deference at all — to the administrator's determinations." 402 F.3d at 74 (citation and internal quotation marks omitted).

We have held that no automatic conflict of interest arises to elevate the standard of review where, as here, "a finding of eligibility means that the insurer will have to pay benefits out of its own pocket." Pari-Fasano v. ITT Hartford Life and Acc. Ins. Co., 230 F.3d 415, 418 (1st Cir.2000). Instead, under such circumstances, courts should apply "the arbitrary and capricious principle, with special emphasis on reasonableness, but with the burden on the claimant to show that the [insurer's] decision was improperly motivated." Id. (citation and internal quotation marks omitted).

Tsoulas asserts two separate arguments to demonstrate that Liberty was improperly motivated. First, she alleges that Liberty undertook the surveillance for the purpose of terminating her benefits. Second, she claims that Liberty engaged in a "selective review" of the administrative record to the exclusion of other evidence.

Tsoulas claims that because there was no change in her condition that could have given rise to suspicion, Liberty must have undertaken the surveillance efforts for the purpose of terminating her benefits. Liberty in turn points out it has made "consistent and continued efforts" to obtain objective evidence of Tsoulas's functional limitations in light of the fact that Tsoulas's own medical providers have repeatedly made contradictory findings regarding her condition. Liberty's claim examiner suggested surveillance as early as September 2000 to "either confirm or call into question" Tsoulas's condition, and Liberty's claim file contains several other references to the need for further information regarding Tsoulas's condition. Further, in its letter to Tsoulas terminating her benefits, Liberty explained that it initiated surveillance when — despite her self-reported need for a cane and her alleged memory and concentration difficulties — she appeared in a January 2004 newspaper photograph standing unaided by a cane and she wrote and submitted an article to the National MS Society website. The district court dismissed Tsoulas's allegation that Liberty initiated surveillance with an improper motive as being "without merit," finding that it was "equally, if not more likely . . . that Liberty initiated surveillance to objectively document whatever her activity level turned out to be." We agree. There is no indication that Liberty undertook the surveillance with the intention of terminating her benefits.

Tsoulas next claims that Liberty engaged in a "selective review" of the medical evidence and the administrative record to the exclusion of other evidence. She claims that Liberty "ignored" the pattern of deterioration and "glossed over" her own doctors' opinions of her disability. This argument is unavailing. The Supreme Court has held that "[n]othing in [ERISA] . . . suggests that plan administrators must accord special deference to the opinions of treating physicians." Black & Decker Disability Plan v. Nord, 538 U.S. 822, 831, 123 S. Ct. 1965, 155 L. Ed. 2d 1034 (2003). Further, it is not "for a court to determine precisely how much weight [an insurer] should have accorded [a particular piece of evidence] in its overall decision." Gannon v. Metro. Life Ins. Co., 360 F.3d 211, 214 (1st Cir. 2004).

Tsoulas next claims that Liberty failed to share with Dr. Erlbaum the supplemental materials she submitted to Liberty after the termination of her benefits. This allegation appears to be totally baseless. In fact, Dr. Erlbaum's report specifically mentions at least two of the supplementary pieces of evidence in question: a brain MRI report from August 2003 and a report from neurologist Dr. T. Edward Collins. The district court inferred from these references that Liberty had sent the entire file to Dr. Erlbaum for review, and we find no error.

Tsoulas also alleges that Liberty itself "did not engage in any substantive review" of that supplemental evidence, an argument that appears similarly meritless. Liberty extended the deadline to permit Tsoulas to submit additional information to support her appeal, at one point noting that "[t]o date [Tsoulas has] forwarded one medical report and several letters from various persons describing [her] physical impairments caused by MS." The record simply does not support her contention that Liberty did not evaluate the supplementary evidence.

As we have previously held, "[t]o affect the standard of review . . . a conflict of interest must be real. A chimerical, imagined, or conjectural conflict will not strip the fiduciary's determination of the deference that otherwise would be due." Leahy v. Raytheon Co., 315 F.3d 11, 16 (1st Cir.2002). We find that Tsoulas has not demonstrated that Liberty was improperly motivated to terminate her benefits, and therefore, that the "arbitrary and capricious" standard of review was appropriate.

Tsoulas next claims that the district court's finding that Liberty's decision to terminate her benefits was not arbitrary, capricious, or an abuse of discretion was flawed because it was not supported by "substantial evidence." Under the arbitrary and capricious standard, "the administrator's decision must be upheld if it is reasoned and supported by substantial evidence." Gannon, 360 F.3d at 213. We find evidence to be substantial "if it is reasonably sufficient to support a conclusion, and the existence of contrary evidence does not, in itself, make the administrator's decision arbitrary." Id.

Tsoulas claims that in assessing her job duties, Liberty improperly relied solely on a Department of Labor Description for Manager, Education and Training based on the Dictionary of Occupational Titles ("DOT"). Under the Plan,

a. If the Covered Person is eligible for the Maximum Own Occupation Benefit, "Disability" or "Disabled" means during the Elimination Period and until the Covered Person reaches the end of the Maximum Benefit Period he is unable to perform all of the material and substantial duties of his occupation on an Active Employment basis because of an Injury or Sickness.

The DOT describes the position of "Manager, Education and Training," as requiring light work, including lifting, carrying, occasional pushing or pulling of 20 pounds, frequent pushing or pulling of lesser weights, frequent reaching, and constant talking. Tsoulas asserts that in defining her occupation Liberty should have acquired a more accurate job description from Medaphis and conducted a more thorough investigation of the requirements of the position. However, Tsoulas bore the "burden to provide evidence that [she] was unable to perform the duties of [her] occupation," and "[a]n integral part of that evidence would be a statement of what [her] job required." Wright, 402 F.3d at 77. Furthermore, although Tsoulas claims that Liberty relied on the DOT description with only "little information" from her, the record reflects that Liberty actually contacted her directly for a description of her job duties. She indicated that her responsibilities included overseeing the department and hiring and firing staff, none of which is inconsistent with the DOT description. In this situation, we adopt the Fourth Circuit's reasoning that "[a] general job description of the DOT, to be applicable, must involve comparable duties but not necessarily every duty." Gallagher v. Reliance Standard Life Ins. Co., 305 F.3d 264, 272 (4th Cir.2002). Tsoulas does not allege that any of her actual occupational requirements were not represented in the DOT description, nor does she identify which, if any, of her job responsibilities she was unable to perform. We find that Liberty did not err in relying on the DOT definition of Manager of Educational Services.

Tsoulas claims the surveillance evidence "falls far short" of establishing that she was able to perform her occupational responsibilities as Manager of Educational Services at Medaphis on a full time basis. To this end, she makes three distinct, but related, arguments. First, she contends that the video is not inconsistent with her self-reported limitations. Second, she insists that even if the video does suggest possible discrepancies, it does not provide a basis for determining that she was capable of returning to work full time. Third, she claims that Liberty acted unreasonably by relying solely upon surveillance evidence in terminating her benefits.

Although Tsoulas reported in her February 12, 2004 Activities Questionnaire that she was unable to walk or stand without an assistive device such as a cane, wheelchair, or scooter, that she could not climb stairs, that her ability to drive a car was "very little," and that she could not go grocery shopping, carry groceries, or put them away, a surveillance video recorded on February 11 (one day prior to her completion of the questionnaire) depicts her standing, walking, and driving without assistance, shopping for groceries, and lifting groceries into her car. Similarly, although she indicated on the questionnaire that she spent 14-18 hours in bed each day, that she "never" left the house on weekends or went shopping at the mall, that she went outdoors only in "summers and fall," and that her daily routine consisted of showering, eating meals, resting, writing emails, and watching television, surveillance videos taken between March 10 and 14 reveal that on those days she drove, stood, and walked unassisted as she went to a bank and a coffee shop and visited a beauty salon. Also during this period, she and her boyfriend drove from Bangor to Portland for a multi-day trip during which she went to restaurants and nightclubs, and to a mall.

Tsoulas first argues that any discrepancies between her questionnaire responses and the video footage is attributable to the ambiguity of the questionnaire. The questionnaire asked "how many hours a day" she walks, and Tsoulas responded "0" to each. In her brief to us she claims that she understood the question to mean "recreational walking" rather than the "minimally necessary walking to get to and fro while engaged in other necessary daily activities." In other words, because Tsoulas never walks more than one hour per day, she claims that her questionnaire response was not directly contradicted by the surveillance. Although we find some degree of logic in this argument, the other inconsistencies remain unexplained.

Tsoulas also alleges that the videos do not contradict her self-reported limitations because in them, she walks "slowly," with "a noticeable limp," and with an "altered gait." She also claims that the video footage of her activities during the surveillance period accounts for only a small — impliedly non-representative — fraction of each day. The district court viewed the video surveillance and found that although "the video depicts Ms. Tsoulas with a stiff-legged gait. . . it is also true the surveillance shows her performing a wide range of activity inconsistent with her stated limitations." Our review of the district court's factual inferences is very deferential, and because we do not find that this conclusion was clearly erroneous, we do not upset it here.

Tsoulas next claims that even if the video footage does contradict her questionnaire responses, the surveillance does not positively support a conclusion that she is capable of performing all of the responsibilities of her job on a full-time basis. Although we agree that Tsoulas's occupational responsibilities differ from the activities portrayed in the surveillance videos, she has not shown that Liberty's inference — that the conflict between her self-reported limitations and her actual activities suggests an ability to perform her occupational responsibilities — was arbitrary and capricious. The DOT description of her employment responsibilities as education manager indicates that her work is typically light or sedentary and although it "sometimes" involves standing, it "almost never" entails walking, running, kneeling, crouching, stooping, or crawling. Similarly, the DOT description indicates that multi-limb coordination and gross body coordination are "not important," and that speed of limb movement is only "somewhat important." We find no error in the district court's determination that Liberty could have rationally concluded that the plaintiff's physical abilities "largely comport with the DOT's description of the physical requirements of her occupation."

Finally, Tsoulas claims that Liberty improperly relied solely upon the surveillance evidence to terminate her benefits. This assertion is plainly unsupported by the record. Liberty consulted two physicians — Dr. Holbrook and Dr. Erlbaum — after reviewing the surveillance videos. Tsoulas gives us no reason to believe that Liberty relied solely upon the surveillance and discounted the medical experts whose advice it sought.

Tsoulas claims that her medical records attest to the severity of her disability and demonstrate that Liberty's termination of her benefits was arbitrary and capricious. To support this argument, she points to abnormalities in her brain MRI scans and the diagnoses and opinions of various physicians, including Dr. Miller's assessment in 1999 that Tsoulas was suffering from relapsing/remitting MS and that she was "greatly disabled" by her symptoms, and Dr. Finder's conclusion in the same year that he did not expect Tsoulas ever to return to either modified or full function.

None of this is sufficient to demonstrate that Liberty lacked substantial evidence to terminate her benefits. The plaintiff's medical history is replete with suggestions that her MS was not totally disabling under the terms of the Plan. In his March 27, 2004 report, Dr. Holbrook, an independent reviewing physician, concluded that the information in Tsoulas's file — including her medical history and the surveillance materials — suggested that Tsoulas had recovered sufficient functional ability to resume full-time work. Liberty then sought a second opinion from Dr. Erlbaum, a board-certified neurologist, who opined that Tsoulas's primary diagnosis was likely psychiatric in nature and that her secondary diagnosis was a mild form of MS. Dr. Erlbaum agreed that Tsoulas was capable of returning to work on a full time basis. Even apart from the fact that the evaluations Tsoulas cites were conducted years before the termination of her benefits, we have held that the existence of conflicting medical evidence is not sufficient to render arbitrary the decision to terminate benefits. Wright, 402 F.3d at 78-79.

Tsoulas also argues that Liberty failed to consider the effects of her fatigue, speech difficulties, and memory problems on her ability to perform her job responsibilities enumerated in the DOT definition. In addition to the physical responsibilities already mentioned, the DOT definition indicated that oral expression and comprehension were "very important" and that memorization, written expression, and comprehension were "important."

Although Tsoulas maintains that her cognitive impairments have been "well-documented," this is not the case. All references to cognitive problems pertain to Tsoulas's self-reported symptoms. In 1998, Dr. Langer-Gould described Tsoulas's mental state as "alert" despite her self-reported experience of "slowed mentation." In July 1999, Dr. Miller noted that Tsoulas was "alert and attentive with normal speech and language, and no gross cognitive defects." Dr. Miller also noted that "there [was] a degree of embellishment" of her symptoms. In 2001, Nurse Entrekin reported that Tsoulas "was significantly impaired due to physical limitations as well as [Tsoulas's] own assertions of deteriorating mental status." The record contains no evidence that any of her examining physicians identified or diagnosed cognitive problems of any kind.

There is significant indication throughout Tsoulas's file that she was suffering from psychological and or psychiatric difficulties. However, despite repeated suggestions by her own physicians, she has largely refused to seek assistance for these problems. When Dr. Langer-Gould suggested that Tsoulas seek treatment for her conversion disorder, Tsoulas "vehemently denied that she had any psychological problems." Tsoulas visited Ms. Kart for psychotherapy sessions for only three months and never obtained another mental health evaluation after March 5, 1999. In the letter denying her appeal, Liberty pointed out that despite her possible psychiatric problems, Tsoulas had not presented a psychiatric medical record from which Liberty could determine whether she had a psychiatric impairment that prevented her from performing her occupational responsibilities. We find no basis in Tsoulas's medical records to upset the district court's conclusion that Liberty's termination of her benefits was not arbitrary or capricious.

Finally, Tsoulas alleges that Dr. Holbrook's review demonstrates a bias in favor of terminating her benefits, and therefore that it was inadequate to support a finding that there was substantial evidence that she was capable of returning to full-time work. Specifically, she claims 1) that Dr. Holbrook "placed undue weight on the existence of a conversion disorder diagnosis in 1998 and 1999" in concluding the diagnosis of MS was not conclusively established; 2) that he relied unreasonably on Dr. O'Connor's opinion that Tsoulas had no physical restrictions because Dr. O'Connor was an oncologist and did not treat her for MS; 3) that Dr. Holbrook did not personally review the surveillance footage, and unreasonably relied on the surveillance reports in concluding that Tsoulas's physical capabilities were greater than she admitted; and 4) that Dr. Holbrook's evaluation did not consider her memory, speech, and cognitive limitations when he concluded that she was able to perform the duties of her own occupation.

Plaintiff's arguments are meritless. Although Dr. Holbrook did specifically rely on Tsoulas's 1998 and 1999 medical records in concluding that her symptoms had a psychological or psychiatric component, Tsoulas cannot now complain that such records are out of date because she refused to seek any psychiatric evaluation subsequent to 1999.

Dr. Holbrook's reference to Dr. O'Connor's opinion that there were "no physical restrictions" was only one of several pieces of evidence that Dr. Holbrook used to determine her current level of functionality. Although Dr. O'Connor was an oncologist who treated Tsoulas for breast cancer, not MS, we do not think it improper for Dr. Holbrook to have taken her assessment into account. We have held that "it is not for a court to determine how much weight [a non-examining medical consultant] should have accorded [a particular piece of evidence]." Gannon, 360 F.3d at 214.

Tsoulas's claim that Dr. Holbrook failed personally to review the surveillance footage is unsupported by the record. Liberty specifically requested that he review the surveillance videos, and his report refers to several incidents shown on the videos. From this, the district court quite reasonably concluded that Dr. Holbrook had actually viewed the videos, and we find no error.

Finally, Tsoulas claims that Dr. Holbrook's review "completely ignored" her memory, speech, and cognitive impairments in concluding that she "does retain functional capacity sufficient for full-time work." Dr. Holbrook's report indicates that he reviewed plaintiff's entire file. As discussed previously, Tsoulas's claim is without basis because her file does not contain clear evidence of these symptoms.

After careful consideration, we cannot say that Liberty's decision to terminate Tsoulas's disability benefits was arbitrary or capricious. For the foregoing reasons, the decision of the district court is affirmed.

Affirmed.