MEMORANDUM OPINION
SAVAGE, District Judge.
This case presents a recurring issue in appeals from a denial of Supplemental Security Income (“SSI”) benefits under Title II of the Social Security Act (“Act”), 42 U.S.C. §§ 401-33 — to what extent must an Administrative Law Judge (“ALJ”) consider, treat and explain Global Assessment of Functioning (“GAF”) scores in determining a claimant’s residual functional capacity.1 The Third Circuit Court of Appeals has not issued a precedential opinion holding that an ALJ’s failure to address GAF scores requires remand. However, district courts in this circuit have repeatedly held that it does.
The plaintiff Gilb Rivera requests review of the ALJ’s decision denying him benefits. He contends the ALJ failed to properly evaluate the medical evidence concerning both his mental and physical impairments, and improperly rejected the treating doctors’ opinions in favor of his own.
After an independent review of the administrative record, we conclude that the ALJ failed to explain how he considered [497]*497Rivera’s eight GAF scores, which were between 40-45 and 51 during a nine-month period. As a result, we are unable to determine whether the ALJ’s decision is supported by substantial evidence. Therefore, we shall remand the case to the Commissioner.
Background and Procedural History
On September 15, 2009, Rivera, who was then thirty-nine, applied for SSI benefits, alleging disability as of December 31, 2007, due to, among other things, fibromyalgia, thoracic disc displacement, myofascial pain, osteoarthritis, gout, and obesity. R. at 16, 18, 22-23, 26, 152-53. He did not claim any emotional or mental problems. R. at 152-53. In 2010, Rivera was diagnosed with mood and anxiety disorders. R. at 23. After his SSI application was denied, Rivera timely requested a hearing. R. at 64-65. Following a hearing held on February 22, 2011, at which Rivera was represented by counsel, the ALJ determined that Rivera had severe impairments, including degenerative disc disease, degenerative joint disease, fibromyalgia, osteoarthritis, obesity, an anxiety disorder, and an affective disorder. R.^at 18. He concluded these impairments did not preclude Rivera’s performing sedentary work, but found him limited to simple unskilled work with only occasional contact with the general public. R. at 21. Relying upon GRID Rule 201.28,2 he then determined that there were jobs available in the national economy that Rivera could perform. R. at 26. Accordingly, on March 11, 2011, the ALJ concluded that Martinez was not disabled. R. at 27.
On September 26, 2012, the Appeals Council denied Rivera’s request for review, making the ALJ’s decision final. R. at 1-4. Rivera then filed this action under 42 U.S.C. § 405(g) seeking judicial review of the Commissioner’s decision.
ALJ’s Findings
The ALJ made the following findings in his March 11, 2011 decision:
1. The claimant meets the insured status requirements of the Social Security Act through December 31, 2012.
2. The claimant has not engaged in substantial gainful activity since December 31, 2007, the alleged onset date (20 CFR 404.1571 et seq.).
3. The claimant has the following severe impairments: degenerative disc disease, degenerative joint disease, fibromyalgia, osteoarthritis, obesity, an anxiety disorder and an affective disorder (20 CFR 404.1520(c)).
4. The claimant does not have an impairment or combination of impair-[498]*498merits that meets or medically equals one of the listed impairments in 20 CFR Part 404, Subpart P, Appendix 1 (20 CFR 404.1520(d), 404.1525 and 404.1526).
5. After careful consideration of the entire record, the undersigned finds that the claimant has the residual functional capacity to perform sedentary work as defined in 20 CFR 404.1567(a) except that he is limited to simple unskilled work with only occasional contact with the general public. However, his remaining mental capacities are sufficient to meet the intellectual and emotional demands of at least unskilled, competitive, remunerative work on a sustained basis. In addition, the claimant is capable of understanding, remembering and carrying out simple instructions; making judgments that are commensurate with the functions of unskilled work (i.e., simple work-related decisions); responding appropriately to supervision, co-workers and usual work situations; and dealing with changes in a routine work setting.
6. The claimant is unable to perform any past relevant work (20 CFR 404.1565).
7. The claimant was born on August 8, 1970 and was 37 years old, which is defined as a younger individual age 18-44, on the alleged disability onset date (20 CFR 404.1563).
8. The claimant has at least a high school education and is able to communicate in English (20 CFR 404.1564).
9. Transferability of job skills is not material to the determination of disability because using the Medical-Vocational Rules as a framework supports a finding that the claimant is “not disabled,” whether or not the claimant- has transferable job skills (See SSR 82-41 and 20 CFR Part 404, Subpart P, Appendix 2).
10. Considering the claimant’s age, education, work experience, and residual functional capacity, there are jobs that exist in significant numbers in the national economy that the claimant can perform (20 CFR 404.1569 and 404.1569(a)).
11. The claimant has not been under a disability, as defined in the Social Security Act, from December 31, 2007, through the date of this decision (20 CFR 404.1520(g)).
R. 18-27.
Standard of Review
On judicial review, the court determines whether the Commissioner’s decision is supported by substantial evidence. 42 U.S.C. §§ 405(g), 1383(c)(3); Hartranft v. Apfel, 181 F.3d 358, 360 (3d Cir.1999). Substantial evidence is “ ‘more than a mere scintilla;’ it means ‘such relevant evidence as a reasonable mind might accept as adequate.’ ” Thomas v. Comm’r of Soc. Sec. Admin., 625 F.3d 798, 800 (3d Cir.2010) (quoting Plummer v. Apfel, 186 F.3d 422, 427 (3d Cir.1999)).
To facilitate meaningful judicial review, the ALJ must explain clearly and fully the basis of his decision. Barren Creek Coal Co. v. Witmer, 111 F.3d 352, 356 (3d Cir.1997) (quoting Cotter v. Harris, 642 F.2d 700, 704-05 (3d Cir.1981)). The ALJ must discuss what evidence supports his determination, what evidence he rejected, and his reasons for accepting some evidence while rejecting other evidence. Cotter, 642 F.2d at 705.
[499]*499Sequential Evaluation
To determine whether a claimant is disabled, the ALJ must apply the familiar five-step sequential process prescribed in the Social Security regulations, 20 C.F.R. § 404.1520(a)(4); Phillips v. Astrue, 671 F.3d 699, 701 (8th Cir.2012); Rutherford v. Barnhart, 399 F.3d 546, 551 (3d Cir.2005). In the first four steps, the claimant must make a prima facie showing of disability by demonstrating that he has a severe impairment that prevents him from performing work he has done in the past. Hoopai v. Astrue, 499 F.3d 1071, 1074 (9th Cir.2007). At step one, the claimant must demonstrate that he is not engaged in gainful employment. Id. At step two, the claimant must show that he has a “severe impairment,” defined as “any impairment or combination of impairments which significantly limits [the claimant’s] physical or mental ability to do basic work activities.” 20 C.F.R. §§ 404.1520(c), 416.920(c). At the third step, the ALJ determines whether the claimant’s impairment or impairments are equal to one of the impairments listed in an appendix to the social security regulations. See 20 C.F.R. pt. 404, subpt. P, app. 1 (2014) (the “Listings”). The Commissioner has decided that the listed impairments are so severe that they conclusively render a claimant disabled. 20 C.F.R. § 404.1520(d); Plummer, 186 F.3d at 428. Thus, if the claimant meets his burden at step three by showing that he has a listed impairment or impairments, he is per se disabled and the inquiry ends. 20 C.F.R. §§ 404.1520(a)(4)(iii), 416.920(a)(4)(iü); Rutherford, 399 F.3d at 551.
If the claimant’s impairment does not equal one of the listed impairments, the inquiry proceeds to step four where the claimant must show that the impairment prevents him from performing his past relevant work. Rutherford, 399 F.3d at 551. Once the claimant has established that he cannot return to his previous work, the process moves to the fifth step. Id. There, the Commissioner has the burden “to demonstrate that ‘the claimant-can perform a significant number of other jobs in the national economy.’ ” Hoopai, 499 F.3d at 1074 (quoting Thomas v. Barnhart, 278 F.3d 947, 955 (9th Cir.2002)); see also Kane v. Heckler, 776 F.2d 1130, 1132 (3d Cir.1985) (citing 42 U.S.C. § 423(d)(2)(A) (1982); Rossi v. Califano, 602 F.2d 55, 58 (3d Cir.1979)). If the Commissioner meets that burden, the ALJ must find that the claimant is not disabled. See 20 C.F.R. §§ 404.1520(a)(4)(v), 416.920(a)(4)(v).
In this case, the ALJ found that Rivera met his burden at steps one and two. R. at 18. At step three, he determined that Rivera did not have an impairment which met or equaled one of the impairments identified in the Listings. The ALJ determined that Rivera’s anxiety and affective disorders did not meet or medically equal the diagnostic criteria of Listing 12.04 because at least two of the “paragraph B” criteria were not satisfied as required. R. at 20.3
[501]*501The ALJ concluded that Rivera had no marked restrictions in activities of daily-living or marked difficulties in maintaining social functioning, concentration, persistence or pace. There were no repeated episodes of decompensation. Instead, he found Rivera had only mild restrictions in daily living activities and moderate difficulties in maintaining social functioning, concentration, persistence or pace. The ALJ found that Rivera’s mental impairment did not cause any “marked” limitation satisfying the “paragraph B” criteria. He also found no evidence to establish the presence of the “paragraph C” criteria. R. at 20.
Proceeding to step four, the ALJ determined that Rivera has the residual functional capacity to perform sedentary work limited to simple unskilled work with only occasional contact with the general public. R. at 21. He also found that Rivera’s “remaining mental capacities are sufficient to meet the intellectual and emotional demands of at least unskilled, competitive, remunerative work on a sustained basis.” R. at 21. Finally, he found that Rivera is “capable of understanding, remembering and carrying out simple instructions; making judgments that are commensurate with the functions of unskilled work (i.e., simple work-related decisions); responding appropriately to supervision, co-workers and usual work situations; and dealing with changes in a routine work setting.” R. at 21. Given this finding and considering the claimant’s age, education, work history, and residual functional capacity, the ALJ concluded that a significant number of jobs existed in the economy which Rivera could perform. R. at 26. Thus, he found Rivera was not disabled. R. at 27.
Discussion
Rivera contends the ALJ’s decision is not supported by substantial evidence because the ALJ failed to adequately consider the psychiatric treatment records, including GAF scores and the opinion of a treating psychiatrist. He argues that the ALJ was mistaken in the belief that a GAF score of 50 is inconsistent with the psychiatrist’s opinion that Rivera is markedly limited in many categories affecting his ability to work.4 He also points out that the ALJ did not discuss all of his GAF scores and did not explain their significance.5 According to Rivera, the ALJ improperly substituted his own judgment in place of the medical evidence, particularly of the treating psychiatrist.6
The Commissioner points out that Rivera did not even allege that he had any disabling mental impairments when he applied for benefits.7 He did not report any problems with remembering, understanding, concentrating, completing tasks, following instructions, or paying attention.8 He did not begin treatment with a mental health specialist until over two years after the alleged disability.9 Moreover, Rivera improved -with treatment.10 Therefore, the Commissioner argues that the ALJ did not err in rejecting the treating psychiatrist’s [502]*502opinion because the evidence did not support it.11
The ALJ has a duty to evaluate all relevant evidence in the record. Fargnoli v. Massanari, 247 F.3d 34, 41 (3d Cir.2001); Burnett v. Comm’r of Soc. Sec., 220 F.3d 112, 121 (3d Cir.2000); Cotter, 642 F.2d at 704, 706. He must explain the evidence supporting his findings and the reasons for discounting the evidence he rejects. Diaz v. Comm’r of Soc. Sec., 677 F.3d 500, 505-06 (3d Cir.2009); Cotter, 642 F.2d at 705-06. Otherwise, the reviewing court cannot determine whether significant probative evidence was improperly rejected or ignored. Burnett, 220 F.3d at 121; Cotter, 642 F.2d at 706-07.
The opinions of a treating physician are entitled to substantial and, in some cases, controlling weight. Johnson v. Comm’r of Soc. Sec., 529 F.3d 198, 202 (3d Cir.2008) (quoting Fargnoli 247 F.3d at 43). The treating physician’s opinions should be given “great weight, ‘especially when their opinions reflect expert judgment based on a continuing observation of the patient’s condition over a prolonged period of time.’” Brownawell v. Comm’r of Soc. Sec., 554 F.3d 352, 355 (3d Cir.2008) (quoting Morales v. Apfel, 225 F.3d 310, 317 (3d Cir.2000)). When he rejects the treating physician’s opinion, the ALJ must adequately explain his reasons for doing so. Fargnoli 247 F.3d at 43-44 (citing Burnett, 220 F.3d at 121). An ALJ may not make “ ‘speculative inferences from medical reports,’ ” and may not reject a treating physician’s opinion “due to his or her own credibility judgments, speculation or lay opinion.” Morales, 225 F.3d at 317-18 (citations omitted) (“[T]he ALJ’s eredi-bility judgments ... alone do not carry the day and override the medical opinion of a treating physician that is supported by the record.”). In other words, the ALJ may not substitute his lay opinion for the medical opinion of a treating physician, especially in cases involving mental disabilities. Id. at 319.
In this case, in his initial application for benefits, Rivera only reported physical impairments. He did not mention any disabling mental impairments. R. at 152-53.12 In September 2009, his treating rheumatologist started Rivera on Effexor because she felt that he was “developing symptoms suggestive of depression due to his longstanding myalgias.” R. at 315. On March 2, 2010, Rivera’s primary care physician referred him to the Family Guidance Center. R. at 666-68. In April 2010, Rivera started seeing Dr. Putnam, a psychiatrist. Dr. Putnam’s evaluation dated April 30, 2010, notes as follows:
The patient was alert. He was oriented to person, place, and date. Memory is 3 out of 3 immediate and 3 out of 3 at 5 minutes. He wears sunglasses. He is somewhat irritable: No eye contact throughout much of the interview. Felt his tone was initially angry and irritable. He did soften throughout the interview and became more receptive and responsive to support. His underlying mood is depressed, and affect was very blunted. He acknowledged he had some disorganization in his thinking but no overt delusional ideation. He was afraid of some of the thinking that he had and whether aborted [sic] on hallucinations but denied any during the interview. He had average intellectual functioning. Insight was fair to poor. His judgment [503]*503is limited although formally intact. His similarities for fruits and transportation were normal. There were no abnormal movements on exam, but he is very much in pain and great difficulty getting out of his chair and ambulating.
We discussed overall issues regarding treatment. He contracts for safety. He does not want to harm himself and does not want [to] have impact upon his children, and he does get very depressed and hopeless.
He is to follow up with his neurosurgeon. He clearly cannot function, arid it should be proof for disability simply based on his chronic pain. At this point, he also cannot function from a psychiatric standpoint.
R. at 670-71. Dr. Putnam diagnosed Rivera with mood disorder “with probable Major Depression.” R. at 671. He assigned him a GAF score of 40-45.
Rivera continued treating with Dr. Putnam and Daniel Parenti, a therapist. Five months after Rivera’s initial visit, Dr. Putnam observed that his patient’s affect was blunted and he had poor eye contact. R. at 673. He was not wearing sunglasses. Id. He was sedated and drowsy. Id. Dr. Putnam assigned a GAF score of 51. Id.
On December 7, 2010, although he was “cooperative” and “logical,” Rivera reported feeling anxious, self-conscious and paranoid about people looking at him even when he is alone, useless, frustrated about not being able to work, and “alive but not living.” R. at 661. His mood was depressed and he had “blunted poor eye contact.” Id. Dr. Putnam noted a lower GAF score of 50.
On December 9, 2010, at a session with Parenti, Rivera reported feeling angry, frustrated and victimized. R. at 663. On December 22, 2010, he reported that he had- up and down swings, but he had just started to take increased medication. R. at 664. He said he went hunting and was with friends. Id. There was less complaining about pain and depression. Id.
Three weeks later, on January 13, 2011, Rivera reported that Klonopin had reduced his anxiety, but he became easily frustrated when he could not perform work tasks without pain. R. at 665. During the three sessions with Parenti on December 9 and 22, 2010 and January 13, 2011, Rivera’s GAF score remained at 50.
On January 15, 2011, during a session with Dr. Putnam, Rivera again reported that Klonopin had helped him, but he still felt restless. R. at 676. He advised that the date for his hearing was approaching. Id. Dr. Putnam noted that Rivera’s mood was “depressed” and he was still struggling. Id. The GAF score remained at 50.
On January 19, 2011, Dr. Putnam completed a psychiatric/psychological impairment questionnaire. R. 679-86. He reported that he had been treating bimonthly since April 2010. He diagnosed Rivera with mood and anxiety disorders. He assessed a GAF score of 50. He identified the clinical findings as appetite disturbance with weight change; sleep disturbance; mood disturbance; emotional lability; anhedonia or pervasive loss of interests; psychomotor agitation or retardation; paranoia or inappropriate suspi-ciousness; feelings of guilt/worthlessness; difficulty thinking or concentrating; social withdrawal or isolation; blunt, flat or inappropriate affect; decreased energy; persistent irrational fears; generalized persistent anxiety; and, hostility and irritability. R. at 680. Dr. Putnam listed Rivera’s primary symptoms as pervasive loss of interests, persistent anxiety, mood disturbance, sleep disturbance, sleep withdrawal or isolation, difficulty concentrat[504]*504ing, and blunt inappropriate affect. R. at 681. Based upon his findings, Dr. Putnam opined that Rivera had “marked” to “extreme” limitations as a result of his impairments that prevent him from doing work-related activities on a day-to-day basis in a regular work setting. R. at 682-84.
In considering Rivera’s mental impairments, the ALJ acknowledged that Rivera suffers from anxiety and affective disorders. R. at 18. He noted that Rivera first received treatment on April 30, 2010, at which time he was diagnosed with a mood disorder with probable depression and a panic disorder, then in remission. R. at 23. Pertinent to this appeal, the ALJ found that Rivera’s GAF score in the follow-up treatment sessions ranged from 50 to 60. Id.
The ALJ rejected Dr. Putnam’s opinion with respect to Rivera’s work-related limitations, explaining as follows:
[Tjhe undersigned has considered the opinion of Mark Putnam, M.D. ... [Dr. Putnam] found that the claimant was markedly limited in 12 categories .... This is not consistent with the record as a whole or the treatment notes that, after his initial evaluation, consistently rated the claimant’s global assessment of functioning (GAF) score to be, at least, 50. Accordingly, the undersigned gives it little weight.
R. at 26.
The ALJ did not explain how a GAF score of 50 is inconsistent with Dr. Putnam’s opinion. A GAF score of 50 or below indicates serious symptoms, while a GAF score of 51 through 60 indicates moderate symptoms. See American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders (DSM-TV-TR) 32, 34 (4th ed.2000). Thus, the difference between 50 and 51 is significant&emdash;the difference between serious and moderate symptoms.
The ALJ apparently overlooked or ignored the fact that Rivera had a GAF score of 40-45 in April 2010. Nor did he consider the score of 51 in October 2010 and the lower six scores of 50 in December 2010 through January 2011. He did not explain why he discounted the significance of Rivera’s GAF scores.
A GAF score is a “numerical- summary of a clinician’s judgment of [an] individual’s overall level of functioning .... ” See DSM-IV-TR at 32. Under the Social Security Administration rules, a GAF score is not considered to have a “direct correlation to the severity requirements.” 65 F.ed.Reg. 50746, 50764-65 (Aug. 21, 2000). Nevertheless, the GAF scale constitutes acceptable and reliable medical evidence. See id.; Colon v. Barnhart, 424 F.Supp.2d 805, 812 (E.D.Pa.2006) (Although a claimant’s GAF score does not have a “ ‘direct correlation to the severity requirements,’ ... [the GAF score] remains the scale used by mental health professionals to ‘assess current treatment needs and provide a prognosis.’ [Therefore, the GAF score is] medical evidence ... and must be addressed by an ALJ in making a determination regarding a claimant’s disability.”) (quoting 65 Fed.Reg. 50764-65); Dougherty v. Barnhart, No. 05-5383, 2006 WL 2433792, at *9 (E.D.Pa. Aug. 21, 2006).
Although it has issued three non-precedential opinions, the Third Circuit has yet to address in a precedential opinion whether an ALJ’s' failure to discuss numerous GAF scores requires remand. See Rios v. Comm’r of Soc. Sec., 444 Fed.Appx. 532 (3d Cir.2011); Gilroy v. Astrue, 351 Fed.Appx. 714 (3d Cir.2009); Irizarry v. Barnhart, 233 Fed.Appx. 189 (3d Cir.2007). The district courts in the Third Circuit have repeatedly held that the ALJ’s failure to specifically discuss a GAF [505]*505score that supports serious impairments in social or occupational functioning is cause for remand. See, e.g., West v. Astrue, No. 09-2650, 2010 WL 1659712, at *4-6 (E.D.Pa. Apr. 26, 2010) (Baylson, J.) (remanding for failure to consider GAF scores and citing seven district court cases from 2004 through 2009 taking the same approach); Sweeney v. Comm’r of Soc. Sec., 847 F.Supp.2d 797, 805 (W.D.Pa.2012); Metz v. Astrue, No. 10-388, 2010 WL 3719075, at *14 (W.D.Pa. Sept. 17, 2010) (ALJ’s determination not supported by substantial evidence where ALJ “did not mention any GAF scores at all- and provided no rationale for rejection of this evidence.”); Wiggers v. Astrue, No. 09-86, 2010 WL 1904015, at *8-9 (W.D.Pa. May 10, 2010) (GAF scores constitute acceptable medical evidence that must be addressed by an ALJ in making a determination regarding a claimant’s disability); Pounds v. Astrue, 772 F.Supp.2d 713, 726 (W.D.Pa.2011); Bonani v. Astrue, No. 10-0329, 2010 WL 5481551, at *6-7 (W.D.Pa. Oct. 15, 2010), report and recommendation adopted, 2011 WL 9816 (W.D.Pa. Jan. 3, 2011); Lust v. Comm’r of Soc. Sec., No. 10-261, 2010 WL 2773205, at *5 (W.D.Pa. July 13, 2010); Burkett v. Astrue, No. 09-26, 2010 WL 724509, at *9 (W.D.Pa. Feb. 26, 2010); Glover v. Astrue, No. 07-2601, 2008 WL 517229, at *1-2 (E.D.Pa. Feb. 27, 2008); Holmes v. Barnhart, No. 04-5765, 2007 WL 951637, at *11 (E.D.Pa. Mar. 26, 2007) (remand required because of ALJ’s failure to acknowledge claimant’s GAF score of 50); Colon, 424 F.Supp.2d at 813-14; Dougherty, 2006 WL 2433792, at *10; Span ex rel. R.C. v. Barnhart, No. 02-7399, 2004 WL 1535768, at *6-7 (E.D.Pa. May 21, 2004) (concluding- that it is not sufficient for an ALJ to mention GAF scores without adequately explaining how or why they were discounted); Escardille v. Barnhart, 2003 WL 21499999, at *6-7 (E.D.Pa. June 24, 2003) (remand required because ALJ opinion did not reveal he seriously considered claimant’s GAF score of 50). In other words, in explaining the rationale for denying disability, the ALJ must demonstrate that he seriously considered and weighed the importance of the GAF scores. See Colon, 424 F.Supp.2d at 813; Span, 2004 WL 1535768, at *4, *6, *7. If the ALJ discounts the GAF score, he must specify his reasons for doing so. Dougherty, 2006 WL 2433792, at *9-10; Span, 2004 WL 1535768, at *6-8 (citing Adorno v. Shalala, 40 F.3d 43, 48 (3d Cir.1994)); see Diaz, 577 F.3d at 505-06; Schaudeck v. Comm’r of Soc. Sec., 181 F.3d 429, 435 (3d Cir.1999) (“[wjhere competent evidence supports a claimant’s claims, the ALJ must explicitly weigh the evidence....”) (citation omitted). Moreover, the ALJ may not “cherry-pick” higher GAF scores in his analysis and ignore GAF scores that may support a disability. See Colon, 424 F.Supp.2d at 813-14; Dougherty, 2006 WL 2433792, at *10 n. 4.
In West, Judge Baylson reviewed several cases in this district which' discuss the issue of whether remand is required where an ALJ fails to address or examine GAF scores of 50 or below. Having reviewed these cases, Judge Baylson concluded:
Because a GAF score constitutes medical evidence accepted and relied upon by a medical source, it should be addressed by an ALJ in making a determination regarding a claimant’s disability. Clearly, the five GAF scores of 50 or below received by Plaintiff indicate serious symptoms. Yet, after examining the record and the GAF scores contained therein, the Court finds that while the ALJ provided an explanation regarding the evidence upon which she relied, the ALJ failed to disclose any reasons for not considering the five GAF scores of 50 or below received by Plaintiff. For this reason, the Court is unable to conclude that the ALJ’s dis[506]*506ability determination is supported by substantial evidence, and remands the case for consideration of plaintiffs GAF scores in conjunction with the other mental health evidence in the record and their effect on her [residual functional capacity]. The Court makes clear that it does not find that the GAF scores in question necessarily indicate that Plaintiff is “disabled” under the Act. Instead, the Court merely requires that the ALJ, on remand, address the multiple GAF scores received by Plaintiff of 50 or below.
2010 WL 1659712, at *6.13
In this case, the ALJ did not meaningfully address the GAF scores. He did not mention the GAF score below 50, specifically in the 40-45 range. Beyond his statement that Rivera had a score of “at least, 50,” he did not address the six GAF scores of 50. Furthermore, he did not explain how the treatment notes were inconsistent with those scores. Nor did he discuss the GAF score that reflected Dr. Putnam’s opinion that Rivera’s symptoms were severe.
The ALJ accorded little weight to Dr. Putnam’s opinion because he concluded that it was “not consistent” with the doctor’s treatment records “as a whole” and was inconsistent with Rivera’s GAF score of “at least, 50.” R. at 26. The ALJ did not explain how Dr. Putnam’s opinion was inconsistent with the record. Dr. Putnam’s entries and Parenti’s progress notes14 contain various notations which raise serious questions about Rivera’s continuing functional problems. See, e.g., “depressed,” “agoraphobic,” R. at 661, “angry,” “frustrated,” R. 663, “having up and down swings,” R. 664, “gets forgetful,” “poor concentration,” “very depressed and sometimes he wishes his life would end.” R. at 669.
The ALJ may have believed that Rivera’s various activities belied the limitations in Dr. Putnam’s opinions. But, he did not point out these inconsistencies. We cannot speculate why he failed to acknowledge these low scores or discounted their importance. See Fargnoli, 247 F.3d at 44 n. 7 (quoting SEC v. Chenery Corp., 318 U.S. 80, 87, 63 S.Ct. 454, 87 L.Ed. 626 (1943)) (“[t]he grounds upon which an administrative order must be judged are those upon which the record discloses that its action was based.”). Our review of the ALJ’s decision is confined to the record upon which it was based. Chenery, 318 U.S. at 87, 63 S.Ct. 454; Fargnoli 247 F.3d at 44 n. 7. The district court cannot substitute its own judgment or analysis to save an inadequate determination. Fargnoli, 247 F.3d at 44 n. 7. In short, the ALJ’s decision must stand or fall on its own.
A failure to discuss GAF scores does not necessarily constitute error where the ALJ conducts a thorough analysis of the medi[507]*507cal evidence regarding plaintiff’s mental impairments. See Coy v. Astrue, No. 08-1372, 2009 WL 2043491, at *13-14 (W.D.Pa. July 8, 2009). The SSA has expressly declined to endorse the GAF scale, finding that it “does not have a direct correlation to the severity requirements of the Social Security mental disorder listings.” Gilroy, 351 Fed.Appx. at 715 (citing 65 Fed.Reg. 50764-65). Like other evidence, a GAF score may be accorded little or no weight depending upon its consistency with the other relevant evidence in the record. Torres v. Barnhart, 139 Fed.Appx. 411, 415 (3d Cir.2005). Here, the ALJ relied on the GAF score to reject Dr. Putnam’s assessment, stating a GAF score of “at least, 50” was inconsistent with his assessment.
The ALJ was free to give less weight to the GAF scores or Dr. Putnam’s opinion. However, if he did, he was obligated to acknowledge the existence of low GAF scores and to explain why they were inconsistent with the evidence in the record. Because he failed to explicitly consider and explain the weight given to all the medical evidence in the record, the ALJ’s conclusions are not supported by substantial evidence.15 See Gober v. Matthews, 574 F.2d 772, 776 n. 15 (3d Cir.1978) (quoting Arnold v. Sec’y of Health, Educ. and Welfare, 567 F.2d 258, 259 (4th Cir.1977)); see also Burnett, 220 F.3d at 121; Plummer, 186 F.3d at 429; Cotter, 642 F.2d at 705. Absent this analysis, we are unable to determine whether the ALJ discredited the GAF scores or simply ignored them. See Metz, 2010 WL 3719075, at *14. Therefore, we shall remand the case, pursuant to the fourth sentence of 42 U.S.C. § 405(g), to address the multiple GAF scores in relation to the other evidence in the case and Rivera’s residual functional capacity.16
JUDGMENT ORDER
AND NOW, this 27th day of March, 2014, in accordance with the Order dated March 27, 2014, remanding this case to the Commissioner (Document No. 21), it is ORDERED that JUDGMENT IS ENTERED in favor • of the plaintiff and against the defendant pursuant to Fed. R.CrvP. 58.