The relationship between potentially traumatic or stressful events, HIV infection and neurocognitive impairment (NCI): a systematic review of observational epidemiological studies

Figures & data

Table 1. Study inclusion and exclusion criteria.

Inclusion criteria	Exclusion criteria
Studies investigating neurocognitive impairment in the context of HIV and potentially traumatic or stressful events, specifically the combined impact of trauma on cognitive impairment in the context of HIV	Studies not investigating the impact of potentially traumatic or stressful events on cognitive impairment in the context of HIV
Adult samples (≥18)	Youth samples (<18)
Observational studies	Randomised controlled trials, quasi-experimental study designs and qualitative studies.
English language studies only	Non-English language studies
NCI measured by a full neuropsychological battery or self-reported measures of NCI that included z-scores based on published normative data	Self-reported measures of NCI that did not include z-scores based on published normative data Screening instruments of cognitive impairment (e.g. IHDS; HDS; MoCA, etc.)

Figure 1. PRISMA flow diagram of search procedure.

Table 2. Quality assessment of papers included in the systematic review (SAQOR tool).

Study	Sample	Country	Control/comparison group	Exposure/outcome measures	Distorting influences	Reporting of data	Overall study quality
Clark et al., 2012	Sample recruited 49 HIV+ and 47- individuals with ELS to explore effects of HIV and ELS on Amygdala morphometry and NCI. In summary, sample is clear.	USA	Comparison group (47 HIV- individuals) included, easily identifiable, source of comparison group clear. In summary, comparison group is clear.	Main exposure: Early Life Stress Questionnaire Main outcome: sMRI and assessments of psychomotor speed, executive functions, verbal memory and fine motor dexterity. In summary, quality of exposure/outcome measurement is adequate	Confounders (demographic variables that differed significantly between groups). In summary, description of influences are unclear.	Missing data Missing data not reported. In summary, reporting of data is unclear. However, data are clearly and accurately presented with p values reported. No confidence intervals reported. In summary, reporting of data is adequate.	Moderate
Clark et al., 2018	Sample recruited 44 HIV+ participants to explore changes in brain volume affected by ELS. Source, method and entry criteria clearly stated. In summary, sample is adequate.	USA	Comparison group included, easily identifiable, source of comparisons explained, comparisons matched, differences between cases and comparisons controlled for. In summary, comparison group is adequate.	Main exposure: Early Life Stress Questionnaire (ELSQ) Main outcome: Reaction Time Task Structural MRI Self-reported cognitive function through The HIV Medical Outcomes Survey (MOS-HIV). In summary, quality of exposure/outcome measurement is adequate.	Confounders (neuropsychiatric symptoms and scanner type) adequately controlled for. In summary, description of influences are adequate.	Missing data not reported. However, data are clearly and accurately presented with p values reported. No confidence intervals reported. In summary, reporting of data is adequate.	High
Deiss et al., 2019	Cohort recruited 200 military beneficiaries living with HIV to explore the relationship between NCI and mental health disorders. Source of the cohort is unclear. Method and entry criteria clearly stated. In summary, cohort is unclear.	USA	Comparison group included (individuals without lifetime history of PTSD), easily identifiable, source of comparison group unclear. In summary, comparison group is unclear.	Main exposure: Composite International Diagnostic Interview. (CIDI) Main outcome: NCI battery measuring verbal fluency, attention/working memory, visuospatial functioning, information processing, learning/recall, executive functions, motor speed and dexterity and effort.). In summary, quality of exposure/outcome measurement is adequate.	Confounders (age, ethnicity, education) adequately controlled for. Potential confounders including cumulative traumas, war deployed to and region of residence not described. In summary, description of influences are unclear.	Missing data not reported. However, data are clearly and accurately presented with p values and confidence intervals reported. In summary, reporting of data is adequate.	Low
Kapetanovic et al., 2020	187 individuals (102 with HIV and IPT; 35 with HIV only; 24 HIV-negative with IPT; 26 HIV-negative). Source of the cohort is clear. Method and entry criteria clearly stated. In summary, sample is adequate.	USA	Comparison group included (HIV-positive with no IPT and HIV-negative with and without IPT), easily identifiable, source of comparisons explained, comparisons matched, differences between cases and comparisons controlled for. In summary, comparison group is adequate.	Main exposure: Client Diagnostic Questionnaire (CDQ). Main outcome: Structural MRI and assessments of attention/working memory, psychomotor speed, executive functions, information processing, verbal fluency, learning, memory, and adult reading In summary, quality of exposure/outcome measurement is adequate.	Confounders (group status, sex, age, race, and education) were adjusted for. Already adjusted T-scores were applied. Models were also adjusted for WTAR (adult reading) scores. In summary, description of influences are clear.	Data are clearly and accurately presented with p values and confidence intervals reported. In summary, reporting of data is adequate.	High
Lin et al., 2011	964 HIV+ individuals with TBI. Source, method and entry criteria clearly stated. In summary, sample is adequate.	USA	Comparison group included (individuals without TBI), easily identifiable, source of comparison group clear. In summary, comparison group is clear.	Main exposure: self-reported medical history. Main outcome: NCI battery measuring overall vocabulary, verbal fluency, attention/working memory, learning, information processing, learning/recall, executive functions, motor speed and dexterity). In summary, quality of exposure/outcome measurement is adequate.	Confounders (age, sex, education, and ethnicity) adequately controlled for. However, number of TBIs and type of TBI were not controlled for. In summary, description of influences are adequate.	Missing data adequately reported.	Moderate
Malan-Muller et al., 2013	The sample recruited 128 women.	South Africa	The sample included n = 83 who were HIV positive.	Main exposure: childhood trauma measured by the childhood trauma questionnaire (CTQ) measuring 28 experiences of childhood trauma. Main outcome: neurocognitive testing using the International Neuropsychological Test Battery developed by the HIV Neurobehavioral Research Centre (HNRC). A research psychologist conducted the testing examining motor, verbal fluency, attention, working memory, speed, learning recall and executive function.	Confounders (age, education, Body Mass Index, trauma sub-type, traumatic life experiences, post-traumatic stress disorder (PTSD), symptomatology and alcohol abuse), LTL, ARV treatment. As all participants were black so ethnicity was not controlled for in the analysis.	Missing data not mentioned.	High
Spies et al., 2012	Sample recruited consisted of 83 (64%) HIV positive and 47 (36%) HIV-negative women.	South Africa	Forty-eight HIV-positive women were exposed to childhood trauma. Among controls, 20 had been exposed to childhood trauma.	Main exposure: childhood trauma measured by the childhood trauma questionnaire (CTQ) measuring 28 experiences of childhood trauma. Main outcome: neurocognitive testing using the International Neuropsychological Test Battery developed by the HIV Neurobehavioral Research Centre (HNRC). A research psychologist conducted the testing examining motor, verbal fluency, attention, working memory, speed, learning recall and executive function.	Age and education corrected z-scores were calculated from all raw NP data and grouped into domains (motor, verbal fluency, working memory, speed, learning, recall, and executive functions).	Missing data not reported. Both statistically and non-statistically significant variables are reported.	High
Spies et al., 2016	A sample of 124 women tested for HIV status were recruited.	South Africa	62 were HIV-positive, 32 with a history of childhood trauma and 30 without, and 62 were HIV-negative, 31 with a history of childhood trauma and 31 without.	Main exposure: childhood trauma measured by the childhood trauma questionnaire (CTQ) measuring 28 experiences of childhood trauma. Main outcome: Neurocognitive function was measured with 10 tests that have been used commonly in HIV research and cover seven ability domains (learning, delayed recall, processing speed, attention/working memory, executive function, verbal fluency, motor ability).	Age, marital status, ethnicity, employment status, years of education and estimated intracranial vault, brain volumes were captured. Estimated intracranial vault was added as a covariate to control for individual differences in head size.	Missing data not reported. Both statistically and non-statistically significant variables are reported.	High
Spies et al., 2017	A sample of n = 67 (51.9%) HIV+ and 50(38.8%) HIV− women.	South Africa	Fifty-three HIV+ women and 18 controls were exposed to childhood trauma.	Main exposure: childhood trauma measured by the childhood trauma questionnaire (CTQ) measuring 28 experiences of childhood trauma. Main outcome: Neurocognitive function was measured with 10 tests that have been used commonly in HIV research and cover seven ability domains (learning, delayed recall, processing speed, attention/working memory, executive function, verbal fluency, motor ability)	Age, marital status, ethnicity, years of education and employment status were captured. A general physical examination was conducted on all participants. Virologic markers of disease progression (CD4 lymphocyte count and viral loads) were captured.	Missing data not reported.	High
Pukay-Martin et al., 2003	251 HIV+ and 82 HIV- gay or bisexual men. Source clearly stated. Method and entry criteria unclear. In summary, sample is unclear.	USA	Comparison group included (HIV- men), easily identifiable, source of comparison group clear. In summary, comparison group is clear.	Main exposure: Psychiatric Epidemiology Research Interview (PERI) Life Events Scale, measuring 47 traumatic events. Main outcome: Wechsler Adult Intelligence Scale-Revised, Selective Reminding Test, Visual. Memory Span Forward and Backward from the Wechsler Memory Scale-Revised, Verbal Concept Attainment Test, Wisconsin Card Sorting Test, Verbal Fluency, Figural Fluency, Trail Making A and B, Grooved Pegboard, and the Paced Auditory Serial Addition Test. In summary, quality of exposure/outcome measurement is adequate.	Confounders (impact of anxiety, depression, age and education) controlled for. In summary, description of influences are adequate.	Missing data and data analysis procedures not reported. However, data are clearly and accurately presented with p values reported. No confidence intervals reported. In summary, reporting of data is adequate.	High
Rubin et al., 2015	Sample recruited 1009 HIV-infected and 496 at-risk HIV-uninfected women to explore the association between perceived stress and verbal memory from the Women’s Inter-Agency HIV Study (WIHS). In summary, sample is clear.	USA	Comparison group included 496 at-risk HIV-uninfected women, easily identifiable, source of comparison group clear. In summary, comparison group is clear.	Main exposure: Perceived Stress Scale (PSS-10) Main outcome: verbal memory performance using the Hopkins Verbal Learning Test (HVLT). Other domains assessed included verbal learning, attention and concentration, executive functioning (behavioural inhibition, mental flexibility, working memory), psychomotor speed, verbal fluency, and fine motor skills. In summary, quality of exposure/outcome measurement is adequate. In summary, quality of exposure/outcome measurement is adequate.	Confounders (annual household income; depression; Hepatitis C status; self- reported recent, former, or never use of marijuana, crack, cocaine, and/or heroin and smoking; self-reported recent heavy alcohol use; recent antidepressant use; and study geographic site. Additional HIV-related clinical variables of interest included cART use and self-reported adherence, duration of antiretroviral therapy use, current CD4 count <200 cells/mm3, current viral load (undetectable, <10,000 cp/ml, ≥10,000 cp/ ml), and nadir CD4 count <200 cells/mm3 adequately reported. In summary, description of influences are adequate.	Missing data adequately reported. However, data are clearly and accurately presented with p values reported. No confidence intervals reported. In summary, reporting of data is adequate.	High
Rubin et al., 2016	Sample recruited included 38 HIV+ women from the Women’s Inter-Agency HIV Study (WIHS) to explore the neurobiological factors contributing to stress-related memory impairment. In summary, sample is clear.	USA	Comparison group (HIV-infected low stress) included, easily identifiable, source of comparison group clear. In summary, comparison group is clear.	Main exposure: Perceived Stress Scale (PSS-10) and PTSD Checklist-Civilian Version (PCL-C) Main outcome: The Hopkins Verbal learning Test (HVLT) to measure verbal learning and memory. Additional cognitive domain measured included: attention/concentration, executive functions, psychomotor speed and verbal fluency sMRI to measure brain volumetric abnormalities In summary, quality of exposure/outcome measurement is adequate.	Confounders (age) adequately reported. In summary, description of influences are adequate.	Missing data not reported. In summary, reporting of data is unclear. However, data are clearly and accurately presented with p values reported. No confidence intervals reported. In summary, reporting of data is adequate.	High
Rubin et al., 2017	Sample recruited included 646 HIV+ and 300 HIV- women from the Women’s Inter-Agency HIV Study to explore the association between perceived stress and PTSD in learning, memory, and fluency impairments. In summary, sample is clear.	USA	Comparison group (300 HIV- women) included, easily identifiable, source of comparison group clear. In summary, comparison group is clear.	Main exposure: Perceived Stress Scale (PSS-10) and PTSD Checklist-Civilian version (PCL-C) Main outcome: learning and memory. Other domains measure included attention, executive function, fluency, psychomotor speed, and motor skills In summary, quality of exposure/outcome measurement is adequate.	Confounders (wave of enrolment, self-reported annual household income, harmful alcohol use, smoking status, marijuana use, and crack, cocaine, and/or heroin use, psychiatric antidepressant/antipsychotic medication, and positive Hepatitis C antibody) adequately reported. In summary, description of influences are adequate.	Missing data not reported. In summary, reporting of data is unclear. However, data are clearly and accurately presented with p values reported. No confidence intervals reported. In summary, reporting of data is adequate.	High
Watson et al., 2018	Sample recruited included 122 HIV+ and 95 HIV- individuals from the Multi-Dimensional Successful Ageing among Adults living with HIV study to explore the effects of trauma, economic hardship, and stress on NCI and everyday function. In summary, sample is clear.	USA	Comparison group (95 HIV- individuals) included, easily identifiable, source of comparison group clear. In summary, comparison group is clear.	Main exposure: self-report Women’s Health Initiative (WHI) Life Events Scale Main outcome: executive functioning, learning, memory (delayed recall), working memory, verbal fluency, speed of information processing, and complex motor skills. In summary, quality of exposure/outcome measurement is adequate.	Confounders (gender, years of education, lifetime MDD, lifetime substance use disorder, lifetime alcohol use, race/ethnicity and lifetime cannabis use) adequately reported. In summary, description of influences are adequate.	Missing data not reported. In summary, reporting of data is unclear. However, data are clearly and accurately presented with p values reported. No confidence intervals reported. In summary, reporting of data is adequate.	High
Womersley et al., 2018	Sample recruited 128 HIV positive women between 18 and 65 years of age, who were fluent in isiXhosa, English and/or Afrikaans and with a minimum literacy level of 5^th grade education.	South Africa	The study examines how childhood trauma interacts with ApoE in relation to NCI so comparison is determined by childhood trauma versus no childhood trauma.	Main exposure: childhood trauma measured by the childhood trauma questionnaire (CTQ) measuring 28 experiences of childhood trauma. Main outcome: Neurocognitive function as measured by the Neurobehavioral Research Centre (HNRC) battery. This battery consists of 17 tests and uses culturally adapted for the South African context. Seven domains of cognitive function: motor ability, processing speed, verbal fluency, learning, delayed recall, attention, working memory and executive function were examined.	Confounder included global cognitive scores.	Missing data not reported. However, data are clearly and accurately presented with p values reported.	High

SAQOR – Systematic Appraisal of Quality in Observational Research.

Table 3. Studies included in the comprehensive review of the literature on the association between potentially traumatic or stressful events and cognitive function among PLWHA.

Study	Total	PLWHA n	HIV n	HIV+ comparison group	Country	Male (%)	Age range or mean	Trauma or stress measure	Mental health assessment	Relationship between trauma and NCI in PLWHA
Clark et al., 2012	96	49	47	-	USA	58	45	ELSQ	-	**
Clark et al., 2018	44	44	-	Yes	USA	59	45	ELSQ	-	*
Deiss et al., 2019	189	189	-	Yes	USA	100	28–44	-	CIDI (current and lifetime history of alcohol use disorder, depression, and PTSD)	**
Kapetanovic et al., 2020	187	137	50	Yes	USA	63	50	CDQ trauma inventory	CDQ	*
Lin et al., 2011	964	964	-	Yes	USA	77	44	Self-reported medical history	-	*
Malan-Muller et al., 2013		83	45	-	South Africa	0	30	CTQ-SF	-	**
Spies et al., 2012	128	83	47	-	South Africa	0	30	CTQ-SF	-	*
Spies et al., 2016	124	62	62	-	South Africa	0	30; 18–50	CTQ-SF	-	*
Spies et al., 2017	117	67	50	-	South Africa	0	33; 18–50	CTQ-SF	-	**
Pukay-Martin et al., 2003	333	251	82	-	USA	100	33	PERI Life Events Scale	-	**
Rubin et al., 2015	1499	1003	496	-	USA	0	46; 25–87	-	PSS-10	***
Rubin et al., 2016	38	38	-	Yes	USA	0	44; 27–59	PCL-C	PSS-10	*
Rubin et al., 2017	964	646	300	-	USA	0	45	PCL-C	PSS-10	*
Watson et al., 2019	217	122	95	-	USA	78	51; 35–65	WHI Life Events Scale	-	*
Womersley et al., 2019	128	128	-	Yes	South Africa	0	33; 21–50	CTQ-SF		*

* = p < 0.05; ** = p < 0.01; *** = p < 0.001

CDQ = Client Diagnostic Questionnaire; CIDI = Composite International Diagnostic Interview; CTQ-SF = Childhood Trauma Questionnaire-Short Form; ELSQ = Early Life Stress Questionnaire; PCL-C = PTSD Checklist-Civilian Version; PERI = Psychiatric Epidemiology Research Interview; PSS-10 = Perceived Stress Scale; WHI = Women’s Health Initiative Life Events Scale.

Table 4. HIV disease parameters of included studies.

Study	PLWHA n	% on ARVs	Mean (SD) duration of illness in years		Mean (SD) CD4 T-cell count (cells/mm^3)		Mean (SD) HIV viral load (copies/mL)
Clark et al., 2012	49	84	High ELS (n = 25) 13.04 (7.98)	Low ELS (n = 24) 14.46 (6.60)	High ELS (n = 25) 460.92 (257.35)	Low ELS (n = 24) 560.96 (313.51)	High ELS (n = 25) 20 Undetectable (<75) 4 High (>400)	Low ELS (n = 24) 16 Undetectable (<75) 6 High (>400)
Clark et al., 2018	44	100	High ELS (n = 26) 16.00 (7.22)	Low ELS (n = 18) 14.56 (6.64)	High ELS 608.85 (272.19)	Low ELS 594.28 (298.40)	High ELS 65% <50	Low ELS 53% <50
Deiss et al., 2019	189	66	5 (2.11 median IQR)		542 (413,697 median IQR)		48 copies/mL (48,3282 median IQR)
Kapetanovic et al., 2020	137	100	IPT+ (n = 102) 17.9 (9.0)	IPT- (n = 35) 14.1 (9.7)	IPT+ (n = 102) Nadir CD4 201.5 (177.5)	IPT- (n = 35) Nadir CD4 217.9 (169.1)	IPT+ (n = 102) All were aviremic (<40 copies/mL)	IPT- (n = 35) All were aviremic (<40 copies/mL)
Lin et al., 2011	964	66	nr		TBI+ (n = 110) Nadir CD4 213.3 (188.5)	TBI- (n = 590) Nadir CD4 219.9 (210.7)	TBI+ (n = 110) 65.1% detectable (based on limit of detection of 50)	TBI- (n = 590) 70.1% detectable (based on limit of detection of 50)
Malan-Muller et al., 2013	83	16	nr		nr		nr
Spies et al., 2012	83	16	nr		405 (259.80)		105,169.5 (407,459.5)
Spies et al., 2016	62	49	nr		431.20 (267.75)		16,172.08 (439,037.76)
Spies et al., 2017	67	81	nr		CT+ (n = 53) 442.04 (234.88)	CT- (n = 14) 413.57 (285.02)	CT+ (n = 53) 61,505.04 (174,046.09)	CT- (n = 14) 81,169.93 (185,206.44)
Pukay-Martin et al., 2003	251	nr	nr		49.1% ˃400mm^3; 27.5% ≥ 200mm^3 and ≤ 400 mm^3; 23.4% ˂200mm^3		nr
Rubin et al., 2015	1003	76	nr		High stress (n = 384) Nadir 208 (157) 44% ˃500 40% ≥200 and <500 16% <200	Low stress (n = 619) Nadir 218 (158) 55% ˃500 34% ≥200 and <500 11% <200	High stress (n = 384) 47% undetectable 37% <10,000 16% ≥10,000	Low stress (n = 619) 55% undetectable 32% <10,000 13% ≥10,000
Rubin et al., 2016	38	79	nr		High stress (n = 18) Nadir 287 (177) 28% ˃500 55% ≥200 and <500 17% <200	Low stress (n = 20) Nadir 325 (226) 74% ˃500 16% ≥200 and <500 10% <200	High stress (n = 18) 44% undetectable 28% <10,000 28% ≥10,000	Low stress (n = 20) 70% undetectable 20% <10,000 10% ≥10,000
Rubin et al., 2017	646	77	nr		521 (401 median IQR)		29% <500
Watson et al., 2018	122	96	17.1 (8.7)		633(425, 851 median IQR)		6.6% detectable
Womersley et al., 2018	128	46	nr		nr		nr

nr = not reported; ELS = early life stress, TBI = traumatic brain injury; CT = childhood trauma; IQR = interquartile range

Table 5. Relationship between potentially traumatic or stressful life events and cognitive function among PLWHA.

			Cognitive domains
Study	PLWHA n	On ARVs (%)	EF	FLU	ATT/WM	PS	LRN/MEM	Motor	Global NCI
Clark et al., 2012	49	84				**
Clark et al., 2018	44	100			*
Deiss et al., 2019	189	66							**
Kapetanovic et al., 2020	137	100			*
Lin et al., 2011	964	66	*		*
Malan-Muller et al., 2013	83	16		**
Spies et al., 2012	83	16				*	*
Spies et al., 2016	62	49	*	*	*	*	*	*
Spies et al., 2017	67	81	**	**
Pukay-Martin et al., 2003	251	nr	*	**	**	*		*	**
Rubin et al., 2015	1003	76					***
Rubin et al., 2016	38	79					*
Rubin et al., 2017	646	77		*			**
Watson et al., 2019	122	96	*		*		*
Womersley et al., 2019	128	46			*			*

nr = not reported; * = p < 0.05; ** = p < 0.01; *** = p < 0.001

Cognitive domains: EF = executive function; FLU = Fluency; ATT/WM = attention and working memory; PS = processing speed; LRN/MEM = learning and memory; VS = visuospatial and visuoconstructive functions; NCI = neurocognitive impairment.

Clark, U. S., Cohen, R. A., Sweet, L. H., Gongvatana, A., Devlin, K. N., Hana, G. N., & Tashima, K. T. (2012, July). Effects of HIV and early life stress on amygdala morphometry and neurocognitive function. Journal of the International Neuropsychological Society, 18(4), 657–668.

 Web of Science ®Google Scholar

Clark, U. S., Arce Renteria, M., Hegde, R. R., & Morgello, S. (2018). Early life stress-related elevations in reaction time variability are associated with brain volume reductions in HIV+ adults. Frontiers in Behavioral Neuroscience, 12, 6.

 Web of Science ®Google Scholar

Deiss, R., Campbell, C. J., Watson, C. W.-M., Moore, R. C., Crum-Cianflone, N. F., Wang, X., … Moore, D. J. (2019). Posttraumatic stress disorder and neurocognitive impairment in a U.S. military cohort of persons living with HIV. Psychiatry, 82(3), 228–239.

 PubMed Web of Science ®Google Scholar

Kapetanovic, S., Norato, G., Nair, G., Julnes, P. S., Traino, K. A., Geannopoulos, K., & Nath, A. (2020, March). Effect of HIV and interpersonal trauma on cortical thickness, cognition, and daily functioning. JAIDS Journal of Acquired Immune Deficiency Syndromes. doi:10.1097/QAI.0000000000002358

 Web of Science ®Google Scholar

Lin, K., Taylor, M. J., Heaton, R., Franklin, D., Jernigan, T., Fennema-Notestine, C., … The CHARTER Group, F. (2011, March). Effects of traumatic brain injury on cognitive functioning and cerebral metabolites in HIV-infected individuals. Journal of Clinical and Experimental Neuropsychology, 33(3), 326–334.

 PubMed Web of Science ®Google Scholar

Malan-Muller, S., Hemmings, S. M., Spies, G., Kidd, M., Fennema-Notestine, C., & Seedat, S. (2013). Shorter telomere length - A potential susceptibility factor for HIV-associated neurocognitive impairments in South African women [corrected]. PloS One, 8(3), e58351.

 Web of Science ®Google Scholar

Spies, G., Fennema-Notestine, C., Archibald, S. L., Cherner, M., & Seedat, S. (2012). Neurocognitive deficits in HIV-infected women and victims of childhood trauma. AIDS Care, 24(9), 1126–1135.

 PubMed Web of Science ®Google Scholar

Spies, G., Ahmed-Leitao, F., Fennema-Notestine, C., Cherner, M., & Seedat, S. (2016, April). Effects of HIV and childhood trauma on brain morphometry and neurocognitive function. Journal of NeuroVirology, 22(2), 149–158.

 PubMed Web of Science ®Google Scholar

Spies, G., Fennema-Notestine, C., Cherner, M., & Seedat, S. (2017, January). Changes in cognitive function in women with HIV infection and early life stress. AIDS Care, 29(1), 14–23.

 PubMed Web of Science ®Google Scholar

Pukay-Martin, N. D., Cristiani, S. A., Saveanu, R., & Bornstein, R. A. (2003). The relationship between stressful life events and cognitive function in HIV-infected men. The Journal of Neuropsychiatry and Clinical Neurosciences, 15(4), 436–441.

 PubMed Web of Science ®Google Scholar

Rubin, L. H., Cook, J. A., Weber, K. M., Cohen, M. H., Martin, E., Valcour, V., & Maki, P. M. (2015, August). The association of perceived stress and verbal memory is greater in HIV-infected versus HIV-uninfected women. Journal of NeuroVirology, 21(4), 422–432.

 PubMed Web of Science ®Google Scholar

Rubin, L. H., Meyer, V. J. J., Conant, R., Sundermann, E. E., Wu, M., Weber, K. M., & Maki, P. M. (2016, August). Prefrontal cortical volume loss is associated with stress-related deficits in verbal learning and memory in HIV-infected women. Neurobiology of Disease, 92(Pt B), 166–174.

 Web of Science ®Google Scholar

Rubin, L. H., Cook, J. A., Springer, G., Weber, K. M., Cohen, M. H., Martin, E. M., & Maki, P. M. (2017, November). Perceived and post-traumatic stress are associated with decreased learning, memory, and fluency in HIV-infected women. AIDS, 31(17), 1401–2393.

 Web of Science ®Google Scholar

Rubin, L. H., Wu, M., Sundermann, E. E., Meyer, V. J., Smith, R., Weber, K. M., & Maki, P. M. (2016, December). Elevated stress is associated with prefrontal cortex dysfunction during a verbal memory task in women with HIV. Journal of NeuroVirology, 22(6), 840–851.

 PubMed Web of Science ®Google Scholar

Watson, C. W.-M., Sundermann, E. E., Hussain, M. A., Umlauf, A., Thames, A. D., Moore, R. C., … Moore, D. J. (2019, January). Effects of trauma, economic hardship, and stress on neurocognition and everyday function in HIV. Health Psychology, 38(1), 33–42.

 PubMed Web of Science ®Google Scholar

Womersley, J. S., Spies, G., Seedat, S., & Hemmings, S. M. J. (2019, April). Childhood trauma interacts with ApoE to influence neurocognitive function in women living with HIV. Journal of NeuroVirology, 25(2), 183–193.

 Web of Science ®Google Scholar