Abstract

The emotional processing in human immunodeficiency virus-seropositive individuals (HIV+) has been scarcely studied. We included HIV+ individuals (n = 107) on antiretroviral therapy (≥2 years) who completed 6 facial processing tasks and neurocognitive testing. We compared HIV+ and healthy adult (HA) participants (n = 40) in overall performance of each facial processing task. Multiple logistic regressions were conducted to explore predictors of poorer accuracy in those measures in which HIV+ individuals performed poorer than HA participants. We separately explored the impact of neurocognitive status, antiretroviral regimen, and hepatitis C virus (HCV) coinfection on the tasks performance. We found similar performance in overall facial emotion discrimination, recognition, and recall between HIV+ and HA participants. The HIV+ group had poorer recognition of particular negative emotions. Lower WAIS-III Vocabulary scores and active HCV predicted poorer accuracy in recognition of particular emotions. Our results suggest that permanent damage of emotion-related brain systems might persist despite long-term effective antiretroviral therapy.

Introduction

The ability to recognize emotions from the information expressed on the faces of other people is important for adequate human social behavior (Adolph et al., 2002), and depends on neural systems composed of cortical and subcortical structures (Haxby, Hoffman, & Gobbini, 2000). In addition to cognitive and motor dysfunctions, facial emotion processing deficits have been described in conditions that affect emotion-related brain systems such as Huntington's disease and Parkinson's disease (Dogan et al., 2014; Péron, Dondaine, Le Jeune, Grandjean, & Vérin, 2012).

Human immunodeficiency virus (HIV) primarily affects the integrity of frontostriatal systems (Heaton et al., 1995; Pfefferbaum et al., 2009), and brain dysfunctions caused by HIV trigger neuropsychological deficits (White, Heaton, & Monsch, 1995). Effective antiretroviral therapy (ART) protects against HIV-induced brain damage, such that the incidence of severe cases of neurocognitive impairment has fallen since the introduction of ART (Sacktor et al., 2001). Moreover, HIV-infected (HIV+) patients receiving effective ART that achieve peripheral viral suppression have the lowest frequencies of neurocognitive impairment. However, selective neuropsychological deficits appear to persist in some aviremic HIV+ patients despite current virological control (Ciccarelli et al., 2013; Cysique & Brew, 2011; Garvey, Surendrakumar, & Winston, 2011; González-Baeza et al., 2014).

Traditionally, tests covering cognitive and motor abilities have been applied to detect neuropsychological deficits associated with HIV infection, but brain-related emotional processing has been scarcely studied. Some studies based on alexithymia questionnaires reported a higher frequencies of self-referred difficulty of identifying feelings in HIV+ patients compared with seronegative individuals, and diminished performance in cognitive tasks that are sensitive to fronto-subcortical dysfunctions in those HIV+ patients with alexithymia. The authors suggest that alexithymia deficits may be a consequence of the effect of HIV infection on frontostriatal systems (Bogdanova, Díaz-Santos, & Cronin-Golomb, 2010; McIntosh et al., 2014). Objective neuropsychological tests have been also applied in this population. Overall preservation of facial emotion processing and selective recognition deficits of fear or sadness have been reported in three cohorts of HIV+ individuals (Baldonero et al., 2013; Clark, Cohen, Westbrook, Devlin, & Tashima, 2010; Clark et al., 2015; Lane, Moore, Batchelor, Brew, & Cysique, 2012). None of those studies was conducted in cohorts of all HIV+ patients virologically suppressed. Thus, objective assessments of facial emotion processing in cohorts of clinically different patients might expand current knowledge. Moreover, Lysaker and colleagues (2012) found low performance formulating ideas about the intentions of others in a small sample of HIV+ patients, suggesting that the emotional aspects of social cognition might also be altered.

The main objective of the present study was to determine whether our sample of long-term virologically suppressed HIV+ patients had facial emotion recognition deficits similar to those reported in cohorts of clinically different HIV+ patients (Baldonero et al., 2013; Clark et al., 2010; Lane et al., 2012). Secondary objectives included: to explore deficits in discrimination and memorization of facial expressions, and the effect of medical variables or global neurocognitive status on facial emotion processing in our HIV+ sample. We have used similar tasks to prior studies that involve discrimination and recognition of basic facial emotions. In addition, we applied tasks that require memorization of facial information, and a new recognition task that does not require verbal labeling of emotions. Based on prior outcomes (Baldonero et al., 2013; Clark et al., 2010; Lane et al., 2012), we expected preservation of overall discrimination and recognition of basic facial emotions, but deficits in the recognition of negative emotions such as fear or sadness. Hypothesis about performance in tasks that require memorization of facial expressions and the impact of medical variables such as hepatitis C virus (HCV) coinfection or the type of antiretroviral regimen used on deficits were not established due to the lack of previous results.

Methods

Participants

We included a total of 107 HIV+ and 40 healthy adult (HA) participants. HIV+ participants were recruited from an established cohort designed to compare the neurocognitive impact of two antiretroviral regimens in Hospital Universitario La Paz HIV Unit (Madrid, Spain) (Pérez-Valero et al., 2014). HIV+ participants enrolled in this study completed several facial emotional processing tasks during a standard neuropsychological assessment. At the moment of the evaluation, patients had to be at least 18 years old. All were receiving ART and plasma virologically suppressed—less than 50 HIV RNA copies/mL—for at least 2 years. We excluded HIV+ patients with neurological comorbidities, patients on active treatment with peginterferon for HCV, and those who had finished the treatment during the previous 6 months. Patients coinfected with HCV were not excluded owing to the high prevalence of HIV/HCV coinfection in Spain (González-García et al., 2005). We also excluded global intellectual disabilities, substance abuse during the previous 3 months, alcohol abuse during the previous 6 months, diagnosis of psychotic or mood disorders according to the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV-TR), neuromuscular disorders, visual dysfunctions, or use of psychiatric medications, which could interfere with results of the neurocognitive evaluation. Patients who had mood disorders sometime in their lifetime were required to be in complete remission for at least 6 months. HAs were relatives of patients recruited from a Spanish Neurology Service and professionals from the recruitment center. HAs were excluded if they had global intellectual disabilities, mood disorders or substance use disorders at the time of the assessment, and history of neurological or history of severe psychiatric disorders. We matched HIV+ and HA participants for gender and age.

The local Ethics Committees for Clinical Research of each participant hospital approved the protocol and the procedures. All participants gave written informed consent and completed the protocol of emotional tasks. In addition, HIV+ participants consented to perform a standard venipuncture and to complete a standard neuropsychological assessment.

Materials and Apparatus

Psychopathological screening and standard neuropsychological assessment

Patients were interviewed by an experienced psychologist if they score equal to or above 8 in the depression subscale of the Hospital Anxiety and Depression Scale. Patients who fulfilled criteria for current major depression (DSM-IV criteria) in the structured interviews were excluded and referred to the Psychiatric Service.

All HIV+ participants underwent a comprehensive neuropsychological test battery to determine presence or absence of neurocognitive impairment. The battery tasks have been previously reported (González-Baeza et al., 2014; Pérez-Valero et al., 2013), and involved 14 measures that estimate the functioning in the following seven ability domains: attention and working memory, mental flexibility, verbal fluency, verbal and visual learning, verbal and visual delayed recall, speed of information processing, and fine motor skills. In addition, the Wechsler Adult Intelligence Scale (WAIS-III) Vocabulary subtest was administered, but we did not use this test to estimate neurocognitive impairment. Following Frascati recommendation, neurocognitive impairment was considered when patients performed at least 1 SD below the mean on the normative data in at least two ability domains (Antinori et al., 2007). The best available normative data for our sample at the moment of the correction were used to covert 14 raw test scores to 14 normalized Z scores. To estimate the performance in each ability domain, a Z score was calculated as the mean of two neuropsychological measures.

During the same visit, the psychologist recorded history of substance abuse and relevant sociodemographical data.

Facial emotional processing tasks

All HIV+ and HA participants performed a computerized battery of tests composed of six tasks that require discrimination, recall and recognition of facial emotions. These tasks were adapted from five subsets of the Florida Affect Battery (FAB), which have been used in several clinical studies (Bowers, Blonder, & Heilman, 1991; Bucks & Radford, 2004,; Carvajal, Rubio, Martín, Serrano, & García-Sola, 2009; Milders, Ietswaart, Crawford, & Currie, 2008). Participants were seated at a distance of approximately 50 cm from a 20 in. 1,680 × 1,050 pixels computer screen. Each task contains 10 slides always presented in the same order. A trial was conducted for each task, and neutral, happiness, sadness, anger, and fear expressions were showed in all tasks. Actresses posing in the stimuli were women of Caucasian ethnicity. Stimuli only included full faces with hair, although other features such as neck or clothes were not showed.

Block 1 included two discrimination tasks. Slides in the first task—facial identity discrimination—show simultaneously two photographs of neutral facial expressions. Participants have to decide whether the photographs correspond to the same or different women. Both pictures were of the same woman in half of slides. Slides in the second task—facial affect discrimination—show two photographs of different women. Subjects have to verbally respond whether the two faces depict the same or different emotional expressions. Both women displayed the same emotional expression in half of the slides.

Block 2 included two facial recall tasks. These tasks were adapted from the FAB facial affect matching task used in a previous study (Carvajal et al., 2009). Both tasks showed for 5 s the slide of a woman who expresses a facial emotional expression (neutral, happiness, sadness, anger, and fear). The slide is then replaced by other slide that showed the pictures of five women with different facial expressions. Participants were exposed to the same 10 stimuli in both tasks. In the first task—recall of the model's face—subjects had to identify the woman seen before, regardless of her facial expression. In the second task—recall of facial expression—participants had to identify the face with the same emotional expression seen before, regardless of the identity of the woman. To properly solve both tasks, participants had to select the chosen face.

Block 3 included two recognition tasks. In the first task—facial affect naming—participants had to verbally label the recognized emotion during the presentation of individual faces. A sheet with five labels written was provided to participants (neutral, happiness, sadness, anger, and fear). The second task—facial affect selection—showed slides displaying simultaneously five photographs of the same woman, each of which expressed five different facial emotions. The examiner said verbally an emotion (neutral, happiness, sadness, anger, and fear) and participants had to select the photography, which corresponds with that emotion. Thus, participants did not have to verbally label the recognized emotion.

Medical and laboratory measures

Plasma HIV RNA level and plasma CD4 T-cell counts were obtained by a standard venipuncture the same week of neuropsychological assessment in all HIV+ participants. A physician registered medical variables. Time since HIV diagnosis, time virologically suppressed, time on ART, current antiretroviral regimen, AIDS status, mode of HIV transmission, prior treatment with peginterferon, and current HCV serostatus were captured from a computerized database.

Data analysis

Absolute and relative frequencies for categorical variables and median (interquartile range [IQR]) for continuous variables were used to describe sample characteristics. Parametric and nonparametric tests were conducted depending on whether the variable was normally distributed.

To compare HIV+ and HA participants in overall facial processing, average raw scores by group and block of tasks (discrimination, recall, and recognition) were calculated using a two-factor group task interaction effect with task as repeated factors on the number of correct responses (analysis of variance [ANOVA]).

The nonparametric Mann–Whitney U-test was conducted to determine whether HIV+ patients had worse accuracy in recognition of each basic facial emotion (neutral, happiness, sadness, anger, and fear) compared with the HA group.

To explore the ability of medical factors and neurocognitive status to predict poorer outcomes in the HIV+ sample, multiple logistic regressions were conducted in a stepwise mode using as dependent variables those in which the HIV+ group had a poorer performance than the group of HAs. Dependent variables were dichotomized using all correct answers as a measure of high accuracy, and at least one error as a measure of lower accuracy. Nine covariates were forced to enter in each regression analysis: current CD4 count (as continuous), nadir CD4 count (under 200; over 200), years on ART (as continuous), years with suppressed HIV viremia (as continuous), years since HIV diagnosis (as continuous), hepatitis C serostatus (active; no active), peginterferon use (prior exposed; not prior exposed), presence of neurocognitive impairment (yes; no), and WAIS-III Vocabulary subtest Z scores (as continuous).

To separately explore the impact of neurocognitive impairment, hepatitis C, and antiretroviral regimen on overall performance in each facial processing tasks, we compared average raw scores yielded by t-test for independent samples between HIV+ patients: neurocognitively impaired participants versus nonimpaired, with current active hepatitis C versus nonactive hepatitis C, and on protease inhibitor monotherapy versus triple therapy. All analyses were conducted in the total sample of HIV+ patients (n = 107) except that which compare the effect of type of treatment. This analysis included only patients who maintained the same antiretroviral regimen for at least 2 years (n = 93). Effect sizes were calculated by Cohen's d.

Significance level was established at p < .05, and all analyses were conducted using IBM® SPSS® Statistics 20.

Results

Participants' Characteristics

The majority of HIV+ participants (n = 107) were men, middle-aged, Caucasians, who had been infected through sexual route and with good current immune status. Most of the HIV+ participants did not have active HCV infection and less than a quarter had been previously treated with interferon. Ninety-three HIV+ participants (87%) were receiving the same antiretroviral regimen—triple therapy or protease inhibitor monotherapy—for at least 2 years, and 14 patients (13%) had switched the antiretroviral regimen during the last year due to toxicity, simplification, reintensification, or enrollment in a clinical trial (see Table 1). HA participants (n = 40) were also predominantly middle-aged men (65%). We did not find significant differences between HA and HIV+ participants in average age (t(145) = 0.96, p = .34), average years of education (t(145) = −0.78, p = .43), gender proportion (χ2(1, N = 147) = 2.28, p = .13), or ethnicity (χ2(1, N = 147) = 0.44, p = .51; Table 1).

Table 1.

Baseline characteristics of participants

 HIV+ participants (n = 107) HA participants (n = 40) 
Caucasian, N (%) 95 (88.8) 37 (92.5) 
Age, median (IQR) 47.4 (43–52) 42.5 (34–54.7) 
Years of education, median (IQR) 10 (8–14) 11.5 (6–16) 
Men, N (%) 77 (72) 26 (65) 
AIDS diagnosis, N (%) 71 (66.4)  
Current antiretroviral regimen, N (%) 
 Triple therapy 51 (47.6)  
 Monotherapy 42 (39.3)  
 Switches during the prior year 14 (13.1)  
Mode of HIV transmission, N (%) 
 Men who have sex with men 34 (31.8)  
 Heterosexual 32 (29.9)  
 Intravenous drug use 31 (29.0)  
 Other 10 (9.3)  
Hepatitis C coinfection, N (%) 
 No 58 (54.2)  
 Past 24 (22.2)  
 Active 25 (23.4)  
Prior treatment with interferon, N (%) 22 (20.6)  
Years since HIV diagnosis, median (IQR) 16.7 (11.6–22.5)  
Years on antiretroviral therapy, median (IQR) 13.7 (7.4–16.8)  
Years virologically suppressed, median (IQR) 6.6 (4.5–10.2)  
Current CD4 (cell/mm3), median (IQR) 595.5 (421–711.8)  
Nadir CD4 (cell/mm3), median (IQR) 170 (50–272)  
Neurocognitive impairment, N (%) 26 (24.3)  
 HIV+ participants (n = 107) HA participants (n = 40) 
Caucasian, N (%) 95 (88.8) 37 (92.5) 
Age, median (IQR) 47.4 (43–52) 42.5 (34–54.7) 
Years of education, median (IQR) 10 (8–14) 11.5 (6–16) 
Men, N (%) 77 (72) 26 (65) 
AIDS diagnosis, N (%) 71 (66.4)  
Current antiretroviral regimen, N (%) 
 Triple therapy 51 (47.6)  
 Monotherapy 42 (39.3)  
 Switches during the prior year 14 (13.1)  
Mode of HIV transmission, N (%) 
 Men who have sex with men 34 (31.8)  
 Heterosexual 32 (29.9)  
 Intravenous drug use 31 (29.0)  
 Other 10 (9.3)  
Hepatitis C coinfection, N (%) 
 No 58 (54.2)  
 Past 24 (22.2)  
 Active 25 (23.4)  
Prior treatment with interferon, N (%) 22 (20.6)  
Years since HIV diagnosis, median (IQR) 16.7 (11.6–22.5)  
Years on antiretroviral therapy, median (IQR) 13.7 (7.4–16.8)  
Years virologically suppressed, median (IQR) 6.6 (4.5–10.2)  
Current CD4 (cell/mm3), median (IQR) 595.5 (421–711.8)  
Nadir CD4 (cell/mm3), median (IQR) 170 (50–272)  
Neurocognitive impairment, N (%) 26 (24.3)  

Facial Recognition Processing

Overall performance in facial emotion recognition tasks did not differ between the HIV+ group and the group of HAs. Repeated-measures ANOVA did not show a significant effect of group or interactions between tasks and HIV-infection status. We only found a main significant effect of type of task in the recognition block. Both HA and HIV+ participants had a better performance in the facial affect selection task than in the facial affect naming task (F(1,145) = 112.60, p < .0001). The selection task that does not require verbal labeling of emotions was easier for both groups. Moreover, the effect size of differences between groups was small in both tasks: the facial affect naming (d = 0.24) and the facial affect selection (d = 0.05).

The HIV+ group had significant lower performance than the HA group on the recognition of several basic emotions. They had significantly poorer recognition of sadness in the facial affect naming task (z = −2.43, p = .015), and poorer recognition of sadness (z = −2.52, p = .012), anger (z = −2.09, p = .036), and fear (z = −2.30, p = .021) in the facial affect selection task (Fig. 1).

Fig. 1.

The comparison between HIV+ individuals and HA participants in the percentage of correct recognition responses given for each specific emotion *Significant differences (p < .05) in distribution of correct response were calculated by Mann–Whitney U-test. Y axis represents % of correct response.

Fig. 1.

The comparison between HIV+ individuals and HA participants in the percentage of correct recognition responses given for each specific emotion *Significant differences (p < .05) in distribution of correct response were calculated by Mann–Whitney U-test. Y axis represents % of correct response.

Facial Discrimination and Recall

We did not find differences between HIV+ participants and a group of HAs in overall performance of any facial discrimination or facial recall tasks. No significant effect of group or interactions between task and group were revealed in any of the two repeated-measures ANOVA conducted.

HA and HIV+ participants performed better the facial discrimination task than the facial affect discrimination task (F(1,145) = 107.18, p < .0001). Also, the recall of the model's face task was easier than the recall of facial expressions task for both groups (F(1,145) = 120.23, p < .0001). Tasks that do not require explicit emotional processing were performed better in either discrimination or recall blocks. Effect size of differences between groups was small in all the tasks: facial discrimination (d = 0.02), facial affect discrimination (d = 0.22), recall of the model's face (d = 0.05), and recall of facial expression (d = 0.11).

Discrimination and recall blocks only had a main effect on type of task. HA and HIV+ participants performed better the facial discrimination task than the facial affect discrimination task (F(1,145) = 107.18, p < .0001). Also the recall of the model's face task was easier than the recall of facial expressions task for both groups (F(1,145) = 120.23, p < .0001). Tasks that do not require explicit emotional processing were performed better in either discrimination or memory blocks.

Effect size of differences between groups was small in all the tasks: facial discrimination (d = 0.02), facial affect discrimination (d = 0.22), recall of the model's face (d = 0.05), and recall of facial expression (d = 0.11).

Impaired Facial Emotional Processing Tasks Among HIV+ Participants and the Effect of Medical and Neurocognitive Factors

Multiple logistic regressions were conducted only in those variables in which HIV+ participants had significant lower accuracy than HA. We tested the association of medical variables and neurocognitive status with lower accuracy (at least one recognition error) of sadness recognition in the facial affect naming task, and lower accuracy of sadness, anger, and fear recognition in the facial affect selection task.

Medical factors, neurocognitive impairment status, or WAIS-III Vocabulary subtest were not significantly associated with a lower recognition of sadness in the facial affect naming task (p = .14).

However, lower Vocabulary subtest scores were independently associated with poorer recognition accuracy of sadness and anger in the facial affect selection that does not require to verbally label the recognized emotion. HIV-related medical factors or neurocognitive impairment status did not predict lower accuracy, although to be coinfected with HCV significantly predicted lower recognition of fear in the facial affect selection (Table 2)

Table 2.

Predictors of lower accuracy in the facial affect selection task

 Vocabulary Z score
 
Hepatitis C virus coinfection
 
Exp (β) 95% CI p value Exp (β) 95% CI p value 
Sadness 0.049 0.282–0.085 .011 — — — 
Anger 0.051 0.293–0.880 .016 — — — 
Fear — — — 2.824 1.09–7.34 .033 
 Vocabulary Z score
 
Hepatitis C virus coinfection
 
Exp (β) 95% CI p value Exp (β) 95% CI p value 
Sadness 0.049 0.282–0.085 .011 — — — 
Anger 0.051 0.293–0.880 .016 — — — 
Fear — — — 2.824 1.09–7.34 .033 

Note: Multiple logistic regressions were conducted in a stepwise mode in those measures in which the HIV+ group had a poorer performance. Table shows variables included in the final models that were independently associated with lower accuracy in recognition. — represents not significant values.

Impact of Neurocognitive Status, Type of Antiretroviral Regimen, and Active HCV Infection in the Overall Performance of HIV+ Participants

In the analyses conducted to compare the average raw test scores depending on the global cognitive status, HIV+ participants with NCI had lower performance than those without NCI in both facial recall tasks. Moreover, there was a trend toward poorer performance in patients with NCI in the facial affect naming task (p = .06), although HIV+ patients with NCI had not a lower performance in the facial affect selection task. We did not find differences in overall facial identity discrimination or facial affect discrimination between patients with and without NCI (Table 3).

Table 3.

Facial emotional processing according to neurocognitive status in the HIV+ sample

Facial emotion process Raw score, mean (SD)
 
t value p value Cohen's d 
Neurocognitively unimpaired Neurocognitively impaired 
Discrimination 
 Facial discrimination 9.72 (0.52) 9.65 (0.75) 0.57 0.57 0.01 
 Facial affect discrimination 8.53 (0.96) 8.28 (1.07) 1.13 0.26 0.25 
Recognition 
 Facial model naming 7.38 (1.07) 6.88 (1.42) 1.90 0.06 0.40 
 Facial affect selection 8.86 (1.12) 8.81 (1.20) 0.22 0.83 0.05 
Recall 
 Recall model's face* 9.25 (0.93) 8.31 (1.19) 4.18 0.000 0.88 
 Recall model expression* 7.70 (1.46) 7.38 (1.93) 5.52 0.000 1.48 
Facial emotion process Raw score, mean (SD)
 
t value p value Cohen's d 
Neurocognitively unimpaired Neurocognitively impaired 
Discrimination 
 Facial discrimination 9.72 (0.52) 9.65 (0.75) 0.57 0.57 0.01 
 Facial affect discrimination 8.53 (0.96) 8.28 (1.07) 1.13 0.26 0.25 
Recognition 
 Facial model naming 7.38 (1.07) 6.88 (1.42) 1.90 0.06 0.40 
 Facial affect selection 8.86 (1.12) 8.81 (1.20) 0.22 0.83 0.05 
Recall 
 Recall model's face* 9.25 (0.93) 8.31 (1.19) 4.18 0.000 0.88 
 Recall model expression* 7.70 (1.46) 7.38 (1.93) 5.52 0.000 1.48 

p value <.05

The effect of antiretroviral strategies was only tested in HIV+ participants who maintained the same antiretroviral type of regimen for at least 2 years (n = 93). We did not find significant differences between patients receiving protease inhibitor in monotherapy or as part of standard triple therapy in any of the facial emotional tasks (p > .05). Effect sizes of differences between strategies were also small in all the tasks (d < 0.50). HIV+ participants with active HCV compared with those without active HCV infection had not significant differences of performance or medium/large effect sizes of differences in any of the facial discrimination, memorization, or recognition tasks.

Discussion

In our sample of long-term aviremic HIV+ individuals, including 23% coinfected with HCV and 24.5% with global neurocognitive impairment, we did not find overall facial emotion recognition impairment. Moreover, HIV+ participants had similar accuracy in recognition of happy and neutral expressions than HAs. Despite major similarities, the HIV+ group showed poorer recognition of sadness in the task that requires verbal labeling of emotions, and poorer recognition of sadness, anger, and fear in the selection task that does not require verbal labeling.

The HIV+ sample had no overall deficits in any of the facial discrimination and facial recall tasks. Moreover, none of the markers of HIV progression predicted lower performance in recognition of specific emotions. Only active hepatitis C reached a significant value, as a predictor of lower accuracy of fear recognition in the task that does not requires verbal labeling of emotions. Otherwise, lower WAIS-III Vocabulary subtest scores significantly predicted poorer recognition of sadness and anger in the selection task. Neurocognitive impairment was not a predictor of any of the specific deficit expressed by HIV+ participants. However, those HIV+ with neurocognitive impairment considered as a group performed worse than patients without cognitive impairment in both facial recall tasks, and showed a trend toward significance in the facial recognition that require to verbally label. Conversely, hepatitis C status or type of ART was not determined poorer overall performance in any of the facial emotion processing tasks among the HIV+ participants.

Our results are consistent with prior reports in which HIV+ patients had not substantial abnormalities in overall discrimination and recognition of basic facial emotions, regardless of ART effectiveness (Baldonero et al., 2013; Clark et al., 2010; Lane et al., 2012). In agreement with Lane and colleagues (2012) but in contrast with Clark and colleagues (2010) and Baldonero and colleagues (2013), we found deficits in the recognition of sadness—but not of fear—in a naming test. Our findings of additional recognition deficits for anger and fear in the selection task are not comparable with the findings of previous studies, because, to our knowledge, ours is the first study that included a recognition task in which participants had to select of five facial expression the one that correspond with the emotion said by the examiner. Conversely, participants had to select of five verbal labels the one that corresponds with a visual facial expression in the naming test. Both tasks involve similar processes because participants had to recognize and select an emotional category. We do not know why the HIV+ participants have additional deficits in recognition of fear and anger in the selection task compared with the naming task. We speculate that the additional deficits observed might indicate that a higher sensitivity of the selection task to detect subtle abnormalities of recognition due to this task was easier for both patients and HA participants. Poorer reaction time for fear recognition and poorer discrimination of the intensity of happy faces despite normal overall recognition in a naming task reported by Lane and colleagues (2012) might also suggest subtle alterations. New confirmatory studies and the application of a large variety of recognition tasks are needed to characterize the emotional recognition deficits that can be found in HIV+ patients.

The differences between our results and those of previous studies might be related to the use and efficacy of ART. Both Clark and colleagues (2010) and Baldonero and colleagues (2013) study reported fear recognition deficits in their HIV+ samples. Importantly, not all patients were receiving effective ART in either study. It is possible that fear recognition deficits might be reversible after initiation of ART. In contrast, impaired recognition of sadness and other subtle deficits have been reported in cohorts of HIV+ patients receiving ART, such as that of Lane and colleagues (2012) and ours. Sadness recognition deficits might be associated with damages in emotion-related brain systems, irreversible upon initiation or after a long period of effective ART. Mild persistent neurocognitive and motor dysfunction despite effective ART was also reported in our (González-Baeza et al., 2014; Pérez-Valero et al., 2013) and in other cohorts of HIV+ patients (Cysique & Brew, 2011; Garvey et al., 2011).

Only the WAIS-III Vocabulary subtest scores contributed significantly to predict a lower accuracy in recognition of specific emotions in our HIV+ sample. Lower Vocabulary scores predicted poorer accuracy in recognition of sadness and anger in the selecting task that does not require verbal labeling. The Vocabulary subtest has been previously used to estimate premorbid intelligence quotient (IQ) and cognitive reserve. In the general population, lower Vocabulary subtest scores have been associated with deficient neurocognitive performance (Corral, Rodríguez, Amenedo, Sánchez, & Díaz, 2006). In addition, aviremic HIV+ participants with lower estimated premorbid IQ more frequently had persistent neuropsychological deficits (Cysique & Brew, 2011). Although confirmatory longitudinal studies with more comprehensive measures of cognitive reserve are needed, we hypothesize that a low cognitive reserve might be associated with a higher risk of persistent deficits in recognition of facial emotions in HIV+ patients.

Markers of HIV-disease progression did not independently predict any recognition deficit for a specific emotion in our HIV+ sample. However, active HCV coinfection reached significant values as a predictor of lower accuracy of fear recognition in the facial affect selection task. Fear recognition has been related to amygdala damage and white matter disconnection (Adolphs, Tranel, Damasio, & Damasio, 1994; Philippi, Mehta, Grabowski, Adolphs, & Rudrauf, 2009). HIV/HCV-coinfected individuals showed subcortical structure dysfunctions, especially in the basal ganglia but not in the amygdala (Garvey et al., 2012). An independent effect of HCV coinfection on white matter abnormalities has also been reported (Gongvatana et al., 2011; Jernigan et al., 2011). Therefore, more widespread disconnection of white matter fibers might mediate poorer fear recognition in the selection task in this subgroup of patients. Unfortunately, we cannot compare our data, because, to our knowledge, no previous studies have assessed processing of facial emotion in HCV-infected individuals.

Although we did not find a significant independent effect of neurocognitive impairment status in any of the deficient processes, HIV+ patients with neurocognitive impairment had significantly worse overall performance than patients without neurocognitive impairment in those tasks that require memorization of facial information. This subset of patients also had a trend toward worse performance in the naming recognition task that requires verbal labeling of emotions (p = 0.06), but not in the selection recognition task. The study by Lane and colleagues (2012), which included mostly virologically suppressed patients (98%) with a shorter time on ART than the patients in the present cohort, revealed significantly poorer overall recognition of facial emotion in a verbal labeling task in patients with neurocognitive impairment. The softer statistical effect found in our study might be associated with the benefit of longer virological suppression in processing recognition of facial emotion in patients with neurocognitive impairment. In our opinion, these outcomes suggest that neurocognitive impairment in persons living with HIV may lead to poorer facial emotional processing driven by deficient recall of faces, and recognition of emotional expression. However, specific recognition deficits found in our entire HIV+ sample might occur regardless of neurocognitive impairment.

The deficits in the recognition of negative emotions that we observed for the first time in an HIV+ sample are similar to those reported in patients with disconnection of the inferior fronto-occipital fasciculus, which is located in the temporal and frontal lobes along the anterior forceps and corpus callosum (Philippi et al., 2009). Poorer recognition of sadness in Parkinson disease has also been associated with impaired functioning of the inferior fronto-occipital fasciculus (Baggio et al., 2012). Amygdala activation has been associated with processing of facial expression of fear (Adolphs et al., 1994) and sadness (Adolph et al., 2002; Baggio et al., 2012); activation of the basal ganglia has been associated with recognition of facial anger (Baggio et al., 2012; Calder, Keane, Lawrence, & Manes, 2004). Our test data are consistent with compromised integrity of frontostriatal and other axonal bundles, such as the posterior sectors of the corpus callosum (Pfeffermbaum et al., 2009), and the reductions in subcortical volumes found in HIV+ patients. Specifically, the volumetric reduction of the amygdala, basal ganglia, and corpus callosum described in HIV+ individuals (Ances, Ortega, Vaida, Heaps, & Paul, 2012) might mediate the selective recognition deficits that we found in our selected HIV+ sample.

Due to the cross-sectional character of our study, we acknowledge that our conclusions should be considered preliminary. The neuropsychological tasks we used are limited, and additional measures could have been more sensitive. For example, tasks with reaction time measures such as those conducted by Lane and colleagues (2012) study could have detected additional deficits in our HIV+ sample of long-term virologically suppressed patients. Moreover, we cannot rule out the effects of unmeasured confounding factors in the differences we found between HIV+ and HA participants. Although we excluded major confounders in both groups, differences of assessment between groups might limit the quality of our results. The assessment of neurocognitive impairment in the HA sample would have been helpful in the interpretation of these differences. However, the results of the analyses conducted only in the sample of HIV+ patients are valuable because we applied to these patients an extensive neuropsychological evaluation. Normative studies in the FAB or larger sample sizes in the HA comparison groups would improve the precision of our conclusions. We also included participants with active HCV infection, because we were interested in studying a representative sample of patients in Spain. Therefore, we are aware that to include patients with active HCV might confound conclusions about the role of HIV infection in facial emotion processing. Further studies that compare groups of HCV-monoinfected and HIV-monoinfected patients in measures of emotional processing would help the understanding of the effect of each type of infection. Finally, our results are only applicable to our selected group of HIV+ participants with good immunological status and long term of viral suppression.

New confirmatory longitudinal studies with larger and better-characterized groups of HAs are necessary to confirm our main findings of overall preservation of facial emotion processing, but specific recognition deficits in HIV+ individuals receiving stable ART. In addition, further studies in cohorts of clinically different HIV+ patients might provide new data on the impact of HIV infection on emotional-related brain systems, and their behavioral consequences. More extensive assessments, including reaction time and recognition of additional emotions, could prove useful for detecting persistent facial emotional deficits despite ART. Furthermore, neuroimaging techniques would facilitate identification of the structural damage associated with deficits.

In summary, our study results do not support a general trend toward poor discrimination, recognition, and recall of facial emotion information in aviremic HIV+ patients on effective ART. However, we do confirm the persistence of the sadness recognition deficits previously described in an HIV+ sample on ART, and for the first time, we found recognition deficits in all the negative facial emotions assessed in a sample of HIV+ patients. Although HIV+ patients with neurocognitive impairment had poorer overall recall of faces than those without neurocognitive impairment, neurocognitive status did not predict lower accuracy in recognition of basic facial emotions. In our opinion, the deficits we found suggest that permanent damage of emotion-related brain systems might persist despite long-term effective ART. If future studies confirm that HIV+ patients have emotional processing deficits, the next revision of the current criteria for the diagnosis of HIV-associated neurocognitive disorders (Antinori et al., 2007) should consider the inclusion of the emotional processing as an additional ability domain to be assessed.

Funding

This work was supported by Fondos de Investigaciones Sanitarias, Instituto de Salud Carlos III (grant PI10/00483). IdiPAZ AIDS and infectious diseases investigator group is partially supported by “Red de Investigación en SIDA” (AIDS Research Network) “RD07/0006/2007”). AG-B has a predoctoral fellowship “(FI11/00338)” supported by Fondo de Investigaciones Sanitarias. Insituto de Salúd Carlos III. IP-V follow a Juan Rodes Program “(13/00016)” supported by Fondo de Investigaciones Sanitarias. Insituto de Salúd Carlos III.

Conflict of Interest

None declared.

Acknowledgments

We thank all the volunteers and all healthcare personnel who participated in the study: Blanca Arribas, Isabel Barrientos, Jose Ignacio Bernardino, Juan Miguel Castro, Marta Galvez, Juan González, and Francisco Xavier Zamora. Thanks to Fondo de Investigaciones Sanitarias, Insituto de Salúd Carlos III, and “Red de Investigación en SIDA.”

References

Adolphs
R.
et al
. (
2002
).
Neural systems for recognizing emotion
.
Current Opinion in Neurology
 ,
12
,
169
179
.
Adolphs
R.
,
Tranel
D.
,
Damasio
H.
,
Damasio
A.
(
1994
).
Impaired recognition of emotion in facial expressions following bilateral damage to the human amygdala
.
Nature
 ,
372
,
669
672
.
Ances
B. M.
,
Ortega
M.
,
Vaida
F.
,
Heaps
J.
,
Paul
R.
(
2012
).
Independent effects of HIV, aging, and HAART on brain volumetric measures
.
Journal of Acquired Immune Deficiency Syndromes
 ,
59
,
469
477
.
Antinori
A.
,
Arendt
G.
,
Becker
J. T.
,
Brew
B. J.
,
Byrd
D. A.
,
Cherner
M.
et al
. (
2007
).
Updated research nosology for HIV-associated neurocognitive disorders
.
Neurology
 ,
69
,
1789
1799
.
Baggio
H. C.
,
Segura
B.
,
Ibarretxe-Bilbao
N.
,
Valldeoriola
F.
,
Marti
M. J.
,
Compta
Y.
et al
. (
2012
).
Structural correlates of facial emotion recognition deficits in Parkinson's disease patients
.
Neuropsychologia
 ,
50
,
2121
2128
.
Baldonero
E.
,
Ciccarelli
N.
,
Fabbiani
M.
,
Colafigli
M.
,
Improta
E.
,
D'Avino
A.
et al
. (
2013
).
Evaluation of emotion processing in HIV-infected patients and correlation with cognitive performance
.
BMC Psychology
 ,
1
,
1
7
.
Bogdanova
Y.
,
Díaz-Santos
M.
,
Cronin-Golomb
A.
(
2010
).
Neurocognitive correlates of alexithymia in asymptomatic individuals with HIV
.
Neuropsychologia
 ,
48
,
1295
1304
.
Bowers
D.
,
Blonder
L. X.
,
Heilman
E.
(
1991
).
Florida affect battery
 .
Gainesville, FL
:
University of Florida
.
Bucks
R. S.
,
Radford
S. A.
(
2004
).
Emotion processing in Alzheimer's disease
.
Aging & Mental Health
 ,
8
,
222
232
.
Calder
A. J.
,
Keane
J.
,
Lawrence
A. D.
,
Manes
F.
(
2004
).
Impaired recognition of anger following damage to the ventral striatum
.
Brain
 ,
127
,
1958
1969
.
Carvajal
F.
,
Rubio
S.
,
Martín
P.
,
Serrano
J. M.
,
García-Sola
R.
(
2009
).
Perception and recall of faces and facial expressions following temporal lobectomy
.
Epilepsy & Behavior
 ,
14
,
60
65
.
Ciccarelli
N.
,
Fabbiani
M.
,
Colafigli
M.
,
Trecarichi
E. M.
,
Silveri
M. C.
,
Cauda
R.
et al
. (
2013
).
Revised central nervous system neuropenetration-effectiveness score is associated with cognitive disorders in HIV-infected patients with controlled plasma viraemia
.
Antiviral Therapy
 ,
18
,
153
160
.
Clark
U. S.
,
Cohen
R. A.
,
Westbrook
M. L.
,
Devlin
K. N.
,
Tashima
K. T.
(
2010
).
Facial emotion recognition impairments in individuals with HIV
.
Journal of the International Neuropsychological Society
 ,
16
,
1127
1137
.
Clark
U. S.
,
Waker
K. A.
,
Cohen
R. A.
,
Devlin
K. N.
,
Folkers
A. M.
,
Pina
M. J.
et al
. (
2015
).
Facial emotion recognition impairments are associated with brain volume abnormalities in individuals with HIV
.
Neuropsychologia
 ,
70
,
263
271
.
Corral
M.
,
Rodríguez
M.
,
Amenedo
E.
,
Sánchez
J. L.
,
Díaz
F.
(
2006
).
Cognitive reserve, age, and neuropsychological performance in healthy participants
.
Developmental Neuropsychology
 ,
29
,
479
491
.
Cysique
L. A.
,
Brew
B. J.
(
2011
).
Prevalence of non-confounded HIV-associated neurocognitive impairment in the context of plasma HIV RNA suppression
.
Journal of Neurovirology
 ,
17
,
176
183
.
Dogan
I.
,
Saß
C.
,
Mirzazade
S.
,
Kleiman
A.
,
Werner
C. J.
,
Pohl
A.
et al
. (
2014
).
Neural correlates of impaired emotion processing in manifest Huntington's disease
.
Social Cognitive and Affective Neuroscience
 ,
9
,
671
680
.
Garvey
L.
,
Surendrakumar
V.
,
Winston
A.
(
2011
).
Low rates of neurocognitive impairment are observed in neuro-asymptomatic HIV-infected subjects on effective antiretroviral therapy
.
HIV Clinical Trials
 ,
12
,
333
338
.
Garvey
L. C.
,
Pavese
N.
,
Ramlackhansingh
A.
,
Thomson
E.
,
Allsop
J. M.
,
Politis
M.
et al
. (
2012
).
Acute HCV/HIV coinfection is associated with cognitive dysfunction and cerebral metabolite disturbance, but not increased microglial cell activation
.
PLoS One
 ,
7
,
e38980
.
Gongvatana
A.
,
Cohen
R. A.
,
Correia
S.
,
Devlin
K. N.
,
Miles
J.
,
Kang
H.
et al
. (
2011
).
Clinical contributors to cerebral white matter integrity in HIV-infected individuals
.
Journal of Neurovirology
 ,
17
,
477
486
.
González-Baeza
A.
,
Carvajal
F.
,
Bayón
C.
,
Pérez-Valero
I.
,
Estébanez
M.
,
Bernardino
J. I.
et al
. (
2014
).
Pattern of neurocognitive function in patients receiving boosted protease inhibitor monotherapy: A detailed neuropsychological study
.
Journal of Neurovirology
 ,
20
,
362
370
.
González-García
J. J.
,
Mahillo
B.
,
Hernández
S.
,
Pacheco
R.
,
Diz
S.
,
García
P.
et al
. (
2005
).
Prevalences of hepatitis virus coinfection and indications for chronic hepatitis C virus treatment and liver transplantation in Spanish HIV-infected patients. The GESIDA 29/02 and FIPSE 12185/01 Multicenter Study
.
Enfermedades Infecciosas y Microbiologia Clinica
 ,
23
,
340
348
.
Haxby
J. V.
,
Hoffman
E. A.
,
Gobbini
M. I.
(
2000
).
The distributed human neural system for face perception
.
Trends in Cognitive Sciences
 ,
4
,
223
233
.
Heaton
R. K.
,
Grant
I.
,
Butters
N.
,
White
D. A.
,
Kirson
D.
,
Atkinson
J. H.
et al
. (
1995
).
The HNRC 500-Neuropsychology of HIV infection at different disease stage
.
Journal of International Neuropsychological Society
 ,
1
,
231
251
.
Jernigan
T. L.
,
Archibald
S. L.
,
Fennema-Notestine
C.
,
Taylor
M. J.
,
Theilmann
R. J.
,
Julaton
M. D.
et al
. (
2011
).
Clinical factors related to brain structure in HIV: The CHARTER study
.
Journal of Neurovirology
 ,
17
,
248
257
.
Lane
T. A.
,
Moore
D. M.
,
Batchelor
J.
,
Brew
B. J.
,
Cysique
L. A.
(
2012
).
Facial emotional processing in HIV infection: Relation to neurocognitive and neuropsychiatric status
.
Neuropsychology
 ,
26
,
713
722
.
Lysaker
P. H.
,
Ringer
J. M.
,
Buck
K. D.
,
Grant
M.
,
Olesek
K.
,
Leudtke
B. L.
et al
. (
2012
).
Metacognitive and social cognition deficits in patients with significant psychiatric and medical adversity: A comparison between participants with schizophrenia and a sample of participants who are HIV-positive
.
The Journal of Nervous and Mental Disease
 ,
200
,
130
134
.
McIntosh
R. C.
,
Ironson
G.
,
Antoni
M.
,
Kumar
M.
,
Fletcher
M. A.
,
Schneiderman
N.
(
2014
).
Alexithymia is linked to neurocognitive, psychological, neuroendocrine, and immune dysfunction in persons living with HIV
.
Brain, Behavior, and Immunity
 ,
36
,
165
175
.
Milders
M.
,
Ietswaart
M.
,
Crawford
J. R.
,
Currie
D.
(
2008
).
Social behavior following traumatic brain injury and its association with emotion recognition, understanding of intentions, and cognitive flexibility
.
Journal of the International Neuropsychological Society
 ,
14
,
318
326
.
Pérez-Valero
I.
,
González-Baeza
A.
,
Estébanez
M.
,
Monge
S.
,
Montes-Ramírez
M. L.
,
Bayón
C.
et al
. (
2014
).
A prospective cohort study of neurocognitive function in aviremic HIV-infected patients treated with one or three antiretrovirals
.
Clinical
 
Infectious Disease
 ,
59
,
1627
1634
.
Pérez-Valero
I.
,
González-Baeza
A.
,
Estébanez
M.
,
Montes-Ramírez
M. L.
,
Bayón
C.
,
Pulido
F.
et al
. (
2013
).
Neurocognitive impairment in patients treated with protease inhibitor monotherapy or triple drug antiretroviral therapy
.
PLoS One
 ,
8
,
e69493
.
Péron
J.
,
Dondaine
T.
,
Le Jeune
F.
,
Grandjean
D.
,
Vérin
M.
(
2012
).
Emotional processing in Parkinson's disease: A systematic review.
Movement Disorders
 ,
27
,
186
199
.
Pfefferbaum
A.
,
Rosenbloom
M. J.
,
Rohlfing
T.
,
Kemper
C. A.
,
Deresinski
S.
,
Sullivan
E. V.
(
2009
).
Frontostriatal fiber bundle compromise in HIV infection without dementia
.
AIDS
 ,
23
,
1977
1985
.
Philippi
C. L.
,
Mehta
S.
,
Grabowski
T.
,
Adolphs
R.
,
Rudrauf
D.
(
2009
).
Damage to association fiber tracts impairs recognition of the facial expression of emotion
.
The Journal of Neuroscience
 ,
29
,
15089
15099
.
Sacktor
N.
,
Lyles
R. H.
,
Skolasky
R.
,
Kleeberger
C.
,
Selnes
O. A.
,
Miller
E. N.
et al
. (
2001
).
HIV-associated neurologic disease incidence changes: Multicenter AIDS cohort study, 1990–1998
.
Neurology
 ,
56
,
257
260
.
White
D. A.
,
Heaton
R. K.
,
Monsch
A. U.
(
1995
).
Neuropsychological studies of asymptomatic human immunodeficiency virus-type-1 infected individuals. The HNRC Group. HIV Neurobehavioral Research Center
.
Journal of the International Neuropsychological Society
 ,
1
,
304
315
.