*Supported by a Veterans Affairs Merit Review grant to Eileen M. Martin.
We thank Roxanna Farinpour, Terri Harris, Mary Ellen Menken, Greg Renck, Lisa Sworowski, and Vibha
Sabharwal for data collection and preparation; Gail Snukst-Torbeck and Gerald Nunnally for patient referrals;
Alice Pau for pharmacologic data; and John Lackey for computer programming.
Portions of the data were presented at Neuroscience of HIV-1 Infection: Basic and Clinical Frontiers, August 2-5,
1994, Vancouver, British Columbia.
Address correspondence to: Eileen Martin, Department of Psychiatry (M/C 913), University of Illinois, 912 S.
Wood St., Chicago, IL 60612, USA. E-mail: firstname.lastname@example.org.
Accepted for publication: July 7, 1999.
Journal of Clinical and Experimental Neuropsychology
1999, Vol. 21, No. 5, pp. 730-735
© Swets & Zeitlinger
Reaction Times are Faster in HIV-Seropositive Patients on
A Preliminary Report*
Eileen M. Martin1,3, David L. Pitrak2,3, Richard M. Novak2,3, Kenneth J. Pursell2,3, and Kathleen
Departments of1Psychiatry and Neurology,2Medicine-Section of Infectious Disease, University of Illinois,
Chicago, and3VA-Chicago Health Care System-West Side Division
We evaluated subclinical mental and motor slowing in 142 HIV-seropositive patients without dementia,
using computerized simple and choice reaction time tasks and self-report measures of psychological dis-
tress. Patients on antiretroviral therapy at the time of testing (n = 79) had significantly faster choice reac-
tion times (p < 0.05), indicating faster mental processing speed, than untreated patients (n = 63). These
faster RTs could not be attributed to differences in age, education, risk factors, degree of immunosuppres-
sion, substance abuse history, peripheral neuropathy, or psychological distress. Reaction time tasks should
be investigated further as potential outcome measures in clinical trials, particularly for subjects with few
or no overt cognitive deficits.
Mental slowing is a prominent characteristic of
HIV-related dementia (Navia, Jordan & Price,
1986). Reaction time (RT) tasks developed in
cognitive psychology are sensitive measures of
subclinical HIV-related mental slowing. Perfor-
mance on RT measures reliably discriminates
HIV-seropositive gay men (E.M. Martin et al.,
1992; Miller, Satz & Visscher, 1991; Wilkie et
al., 1990) and injection drug users (E.M. Martin
et al., 1995) without dementia from cohort-
matched seronegative controls. Slowed RT per-
formance correlates with CSF levels of quino-
linic acid (A. Martin et al., 1992), which has
been implicated in the pathophysiology of HIV-
related dementia, and with decline in CD4 lym-
phocyte counts (Bornstein et al., 1990). How-
ever, there are few data available on reaction
time performance of HIV-seropositive patients
as a function of antiretroviral treatment status.
Zidovudine (ZDV) crosses the blood-brain
barrier (Klecker et al., 1987), and neuropsycho-
logical and neurological studies confirm that
behavioral function. Schmitt et al. (1988) first
reported that HIV-seropositive subjects’ neuro-
ZDV therapy. Investigators have since reported
amelioration of symptoms of AIDS dementia
complex in adults (Sidtis et al., 1993) and of
HIV encephalopathy in children (Brouwers et
al., 1990; Pizzo et al., 1988) during ZDV ther-
apy. Patients on ZDV therapy also show im-
REACTION TIMES AND ANTIRETROVIRAL THERAPY
provement in laboratory markers of CNS infec-
tion such as CSF $2 microglobulin concentra-
tion (Brew et al., 1992), although the relation-
ship between antiretroviral therapy and CNS
et al., 1997); and increased cerebral glucose me-
tabolism is apparent on PET studies of patients
receiving ZDV (Brunetti et al., 1989). Fewer
data are available on the neurobehavioral effects
of other antiretroviral compounds such as
didanosine, lamivudine, or the protease inhibi-
tors (but see Ferrando et al., 1998; Portegies et
al., 1994; Sidtis et al., 1997).
The relationship between antiretroviral ther-
apy and performance on RT tasks has not been
widely studied. However, RT tasks are advanta-
geous in that they are more rapidly administered
and often more sensitive to HIV-related cogni-
tive dysfunction than clinical neuropsychologi-
cal tests. In this preliminary report, we present
cross-sectional data from a series of 142 HIV-
ipated in RT testing as part of a larger ongoing
investigation of cognition in HIV-1 infection.
We compared simple and choice RTs of subjects
on antiretroviral therapy with those of untreated
patients. Our data indicate that subjects treated
with antiretrovirals at the time of testing had
faster RTs than untreated subjects.
122 male and 20 female HIV-seropositive patients
without dementia participated in the study. This
cohort had no overlap with the HIV-seropositive
group reported in an earlier publication on simple
and choice RTs (E.M. Martin et al., 1992). The
current sample included 62 persons with a history
of injection drug use (IDU), 75 persons who con-
tracted HIV-1 from unprotected sexual activity,
and 5 persons with unknown risk factors. All sub-
jects were ambulatory and without serious clinical
disease at the time of testing, which was conducted
on an outpatient basis. All subjects were without
dementia by history and on clinical interview and
were capable of giving informed consent. They
were recruited from Infectious Disease clinics at
the VA-West Side and the University of Illinois
nizations. Study subjects were paid volunteers and
were not referred because of neurobehavioral
symptoms or signs. Seventy-one individuals were
clinically asymptomatic or had lymphadenopathy
only (CDC Stage A; Centers for Disease Control,
1992). Seventy-one individuals were clinically
symptomatic. This group included 42 persons at
CDC Stage B, who had constitutional symptoms
such as fatigue and night sweats, minor opportu-
nistic infections such as oral candidiasis (n = 27),
or peripheral neuropathy (n = 15) but no AIDS-
defining opportunistic conditions or neurologic
disorders; and 24 persons with various AIDS-de-
fining opportunistic infections or neoplasms (CDC
Stage C) such as Kaposi sarcoma, Pneumocystis
carinii pneumonia, or pulmonary tuberculosis but
no AIDS-defining neurologic disease. Subjects
with a history of any HIV-related neurologic con-
dition other than neuropathy or aseptic meningitis
were excluded from study. Approximately 36% of
all subjects had AIDS-defining CD4 counts.
became available.) The sample is 77% African-
American, 16% Caucasian, 7% Hispanic and 14%
female. Nineteen subjects are left-handed.
Seventy-nine subjects were on antiretroviral
therapy at the time of testing (see Table 1). The
majority (65, or 82%) was on ZDV monotherapy.
Fourteen persons were treated with didanosine or
zalcitabine alone or in combination with ZDV; the
majority of these 14 individuals had been treated
previously with ZDV monotherapy. Sixty-three
individuals had never been treated with anti-
retrovirals. The treated and untreated groups did
not differ significantly in ethnic composition
(P2(1) = 1.09, p < 0.30), incidence of injection
drug use (P2< 1), mean age (t < 1), mean years of
education (t < 1), or incidence of left-handedness,
psychiatric treatment, closed head injury, alcohol
abuse or current treatment with antidepressants or
anxiolytics (P2< 1 for each comparison). Com-
were more immunosuppressed as indexed by their
significantly lower mean CD4 lymphocyte count
(t(135) = 5.05, p < 0.0001), and they had a signifi-
cantly higher incidence of symptomatic disease
(P2(1) = 6.42, p < 0.02) .
Each subject gave written informed consent after
the experimental procedures were explained. We
administered the RT tasks on an AT-compatible
personal computer and NEC Multisync 3D 14"
VGA monitor with ATI-Wonder graphics card.
These tasks have been described in detail else-
where (E.M. Martin et al., 1992). Briefly, both the
EILEEN M. MARTIN ET AL.
Table 1. Demographic Data and Psychological Distress Scores.
(n = 79)
(n = 63)
Note. BDI = Beck Depression Inventory; STAI = State-Trait Anxiety Inventory.
simple and choice RT procedures required the sub-
ject to press a response key as quickly as possible
following the appearance of a visual target stimu-
lus in the center of the computer screen. At the
start of each RT trial, a visual warning stimulus
consisting of a white square appeared in the center
of the screen for 500 ms followed by a target stim-
ulus 500 or 1500 ms later. The ISI varied randomly
across trials, so the subject could not predict target
onset reliably. The target consisted of either a
green or red square. The target remained on the
screen for 3 s or until the subject responded. A new
trial began 1000 ms later.
In the simple condition, we instructed the sub-
ject to press a single key with the dominant hand
as quickly as possible after the target appeared,
without regard to target color. We ran one block of
16 practice and 32 experimental trials. Choice RT
trials were identical in procedure, except that we
instructed the subject to press one response key
with the dominant hand if a red target appeared
and to press a different key with the nondominant
hand if a green target appeared; thus, the motor
response requirements of the two tasks were simi-
lar but the choice task added a decision compo-
nent. We ran 16 practice and two blocks of 32
choice experimental trials. The computer recorded
RTs to the nearest ms, errors and failures to re-
spond within 3 s. Data from error and no-response
trials were excluded from analysis.
In addition, subjects completed the Beck De-
pression Inventory (BDI; Beck, Ward, Mendelson,
& Erbaugh, 1961) and the State-Trait Anxiety In-
ventory-State version (STAI; Spielberger, Gor-
such, & Lushene, 1970). Scores on these measures
were used to monitor the effects of psychological
distress on RT performance.
Median rather than mean reaction times were ana-
lyzed in order to minimize the effects of outliers.
Reaction times and psychological distress scores
were compared using the t-statistic. Pearson corre-
lations were used to analyze relationships between
reaction times, psychological distress scores, and
CD4 lymphocyte counts. Two-tailed " values of
.05 were employed for all analyses.
Mean error rates were essentially zero for the
simple task and approximately 1.5% for the
choice task. There were no significant differ-
ences between groups in mean error rates for
either task, t < 1 for each comparison.
Figure 1 shows the mean RTs for the two
groups. Mean simple reaction times did not dif-
fer significantly between the groups (t < 1).
However, the treated group had significantly
faster choice RTs compared to the untreated
group (t(139) = 2.02, p < 0.05). There were no
significant differences in simple or choice RTs
between subjects on zidovudine monotherapy
compared with subjects on other antiretroviral
regimens (t < 1 for each comparison).
REACTION TIMES AND ANTIRETROVIRAL THERAPY
Fig. 1.Mean simple and choice reaction times for all subjects. Choice reaction times for subjects on
antiretroviral therapy were significantly faster than choice reaction times for untreated subjects (p <
Table 1 shows mean psychological distress
scores for the two groups, with scores for the
BDI subjective and somatic item clusters pre-
sented separately. STAI and subjective BDI
scores did not differ significantly between
groups (STAI: t(138) = 1.61, p < 0.12; BDI Sub-
jective: t(139) = 1.61, p < 0.12). The treated
group had slightly but significantly lower so-
matic BDI scores compared with the untreated
group, t(139) = –2.27, p < 0.05. Simple RTs
were essentially uncorrelated with STAI or BDI
(total or subscale) scores ( r < 0.07 in each in-
stance). Similarly, choice RTs did not correlate
significantly with any psychological distress
scores ( r < 0.13 in all instances).
There were no significant differences in simple
or choice RTs between groups with or without a
positive history of alcohol abuse, psychiatric
history, injection or non-injection drug use, or
anxiolytic/antidepressant treatment ( t < 1 for
each comparison). Further, RTs for subjects
with peripheral neuropathy (n=14) did not differ
significantly from RTs of subjects with no neu-
ropathy (Simple RTs: t(136)= –1.11, p < 0.27,
ChoiceRTs: t <1).RTswerenonsignificantly
correlated with absolute CD4 count ( r < 0.01
in each instance) or with CD4/CD8 ratio (Sim-
ple RTs: r < 0.01, Choice RTs: r = –.20, p <
In this preliminary cross-sectional study, we
found that HIV-seropositive subjects currently
on antiretroviral therapy had significantly faster
choice reaction times, indicating faster mental
processing speed, compared with untreated sub-
jects, despite more advanced immunosuppres-
sion and greater prevalence of symptomatic dis-
ease in the treated group. The groups were well-
matched on demographic variables, including
age, education, and psychiatric and substance
abuse history and thus the differences in ob-
tained RTs cannot be attributed to differing de-
mographic characteristics for the two groups.
Compared with the untreated subjects, the
treated group reported slightly fewer somatic
symptoms such as sleep or appetite problems,
but these complaints did not correlate with RTs,
suggesting that the group differences in RTs
EILEEN M. MARTIN ET AL.
could not be attributed to nonspecific effects of
differing levels of subjective well-being. Fur-
ther, these latter findings provide no evidence in
support of the hypothesis that differences in RTs
were associated with other psychological vari-
ables (e.g., greater optimism in health beliefs
influencing treatment-seeking), although this
possibility might be helpful to address in later
RTs of subjects on zidovudine monotherapy
did not differ significantly from RTs of subjects
on didanosine or zalcitabine. This finding is of
questionable significance, since most subjects
on ddI orddC were on combination therapy with
ZDV; only 10% of treated patients, in fact, were
not receiving ZDV at the time of RT testing.
Certainly, future studies must evaluate RTs for
patients on differing antiretroviral therapies,
since thesecompounds differ in theircapacity to
cross the blood-brain barrier (Balis, Pizzo, &
Our inferences are necessarily limited by the
conducted with drug-naive subjects pre- and
post-initiation of antiretroviral therapy in order
to draw conclusions about the effect of
antiretrovirals on reaction time performance:
these studies are currently in progress in our
laboratory. Additional data must also be ob-
tained on the relationship between duration and
dosage of antiretroviral therapy and cognition,
since available data indicate that neuro-
ever, the current findings and more recent data
we have obtained using a more complex voice-
activated reaction time task (Martin et al., 1998)
are consistent with results of previous studies
apy on laboratory and clinical indices of CNS
infection by HIV-1 (Schmitt et al., 1988; Sidtis
et al., 1993) and suggest that RT measures have
much potential utility as outcome measures in
trials of newer antiretroviral compounds and
combination therapies. It is possible that some
clinicaltrial subjectswithoutevidence ofsignif-
icant cognitive deficits at study entry might
show improved mental function that is less
readily detected by standard clinical neuropsy-
chological tests. If so, RT measures would pro-
vide a useful adjunct measure of quality of life
in patients without dementia. Indeed, such mea-
sures may prove of particular significance in the
therapy with protease inhibitors.
Balis, F.M., Pizzo, P.A., & Butler, K.M. (1992). Clin-
ical pharmacology of 2 ,3 -dideoxyinosine in hu-
man immunodeficiency virus-infected children.
Journal of Infectious Diseases, 165, 99-104.
Beck, A.T., Ward, C.H., Mendelson, M., & Erbaugh,
J.K. (1961). An inventory for measuring depres-
sion. Archives of General Psychiatry, 4, 561-571.
Bornstein, R.A., Nasrallah, H.A., Para, M.F., Fass,
R.J., Whitacre, C.C., & Rice, R.R. Jr. (1991). Rate
of CD4 decline and neuropsychological perfor-
mance in HIV-1 infection. Archives of Neurology,
Brew, B.J., Bhalla, R.B., Paul, M., Sidtis, J.J., Keilp,
J.J., Sadler, A.E., Gallardo, H., McArthur, J.C.,
Schwartz, M.K., & Price, R.W. (1992). Cerebro-
spinal fluid2-microglobulin in patients with AIDS
dementia complex: an expanded series including
response to zidovudine treatment. AIDS, 6, 461-
Brouwers, P., Moss, H., Wolters, P., Eddy, J., Balis,
F., Poplack, D.G., & Pizzo, P.A. (1990). Effect of
continuous-infusion zidovudine therapy on
neuropsychologic functioning in children with
symptomatic human immunodeficiency virus in-
fection. Journal of Pediatrics, 117, 980-985.
Brunetti, A., Berg, G., Di Chiro, G., Cohen, R.M.,
Yarchoan, R., Pizzo, P.A., Broder, S., Eddy, J.,
Fulham, M.J., Finn, R.D., & Larson, S.M. (1989).
Reversal of brain metabolic abnormalities follow-
ing treatment of AIDS dementia complex with 3 -
azido-2 ,3 -dideoxythymidine (AZT, zidovudine):
a PET-FDG study. Journal of Nuclear Medicine,
Centers for Disease Control. (1992). 1993 revised
classification system for HIV infection and ex-
panded surveillance case definition for AIDS
Weekly Report, 41, 1-19.
Ferrando, S., van Gorp, W., McElhinney, M., Goggin,
K., Sewell, M., & Rabkin, J. (1998). Highly active
antiretroviral treatment in HIV infection: benefits
for neuropsychologicalfunction. AIDS,12,F65-70
Klecker, R.W. Jr., Collins, J.M., Yarchoan, R.,
Thomas, R., Jenkins, J.F., Broder, S., & Myers,
C.E. (1987). Plasma and cerebrospinal fluid phar-
macokinetics of 3 -azido-3 deoxythymidine: a
novel pyrimidine analog with potential application
for the treatment of patients with AIDS and related
REACTION TIMES AND ANTIRETROVIRAL THERAPY
Martin, A., Heyes, M.P., Salazar, A.M., Kampen,
D.L., Williams, J., Law, W.A., Coats, M.E., &
Harkey, S.P. (1992). Progressive slowing of reac-
tion time and increasing cerebrospinal fluid con-
centrations of quinolinic acid in HIV-infected indi-
viduals. Journal of Neuropsychiatry and Clinical
Neurosciences, 4, 86-93.
Martin, E.M., Pitrak, D.L., Pursell, K.J., Andersen,
B.A., Mullane, K.M., & Novak, R.M.. (1998). In-
formation processing and antiretroviral therapy in
HIV-1 infection. Journal of the International Neu-
ropsychological Society, 4, 329-335.
Martin, E.M., Pitrak, D.L., Robertson, L.C., Novak
R.M., Mullane, K.M, & Pursell, K.J. (1995).
Global-local analysis in HIV-1 infection. Neuro-
psychology, 9, 102-109.
Martin, E.M., Sorensen, D.J., Edelstein, H.E., & Rob-
ertson, L.C. (1992). Decision-making speed in
HIV-1 infection. AIDS, 6, 109-113.
McArthur, J.C., McClernon, D.R., Cronin, M.F.,
Nance-Sproson, T.E., Saah, A.J., St. Clair, M., &
Lanier, E.R. (1997). Relationship between human
immunodeficiency virus-associated dementia and
viral load in cerebrospinal fluid and brain. Annals
of Neurology, 42, 689-98.
Miller, E.N., Satz, P., & Visscher, B. (1991). Comput-
erized and conventional neuropsychological as-
sessment of HIV-1 infected men. Neurology, 41,
Navia, B.A., Jordan, B.D., & Price, R.W. (1986). The
AIDS dementia complex: I. Clinical features. An-
nals of Neurology, 19, 517-524.
Pizzo, P.A., Eddy, J., Falloon, J., Balis, F.M., Mur-
phy, R.F., Moss, H., Wolters, P., Brouwers, P.,
Jarosinski, P., Rubin, M., Broder,.S., Yarchoan, R.,
Brunetti, A., Maha, M., Nusinoff-Lehrman, S., &
symptomatic HIV infection. New England Journal
of Medicine, 319, 889-896.
Portegies, P., Enting, R.H., de Jong, M.D., Danner,
S.A., Reiss, P., Goudsmit, J., & Lange, J.M.
(1994). AIDS dementia complex and didanosine.
Lancet, 344, 759.
Schmitt, F.A., Bigley, J.W., McKinnis, R., Logue,
P.E., Evans, R.W., & Drucker, J.L. (1988). Neuro-
psychological outcome of zidovudine (AZT) treat-
ment of patients with AIDS and AIDS-related
complex. New England Journal of Medicine, 319,
Sidtis, J.J., Gatsonis, C., Price, R.W., Singer, E.J.,
Collier, A.C., Richman, D.D., Hirsch, M.S.,
Schaerf, F.W., Fischl, M.A., Kieburtz, K.,
Simpson, D., Koch, M.A., Feinberg, J., Dafni, U.,
and the AIDS Clinical Trials Group (1993).
Zidovudine treatment of the AIDS dementia com-
Neurology, 33, 343-349.
Spielberger, C.D., Gorsuch, R.L., & Lushene, R.E.
(1970). State-trait anxiety inventory. Palo Alto,
CA: Consulting Psychologist Press.
Wilkie, F.L., Eisdorfer, C., Morgan, R., Loewenstein,
D.A., & Szapocznik, J. (1990). Cognition in early
of Neurology, 47, 433-440.
Yarchoan, R., Thomas, R.V., Grafman, J., Wichman,
A., Dalakas, M., McAtee, N., Berg, G., Fischl, M.,
Perno, C.F., Klecker, R.W., et al. (1988). Long-
term administration of 3 -azido-2 ,3 -dideoxythy-
midine to patients with AIDS-related neurological
disease. Annals of Neurology, 23 (suppl), S82-S87.