Content uploaded by Sherry H Stewart
Author content
All content in this area was uploaded by Sherry H Stewart
Content may be subject to copyright.
&p.1:Abstract Until recently, research indicated that all ben-
zodiazepines impair explicit memory, while only loraze-
pam impairs priming. Stewart and associates provided
preliminary data which indicated that both oxazepam and
lorazepam may impair implicit memory, but in a time-de-
pendent fashion. The present study was designed to repli-
cate Stewart et al.’s findings after overcoming several lim-
itations of the original study. Thirty subjects were admin-
istered an acute dose of lorazepam (2 mg), oxazepam
(30 mg) or a placebo and were tested with an implicit
(word-stem completion) test and an explicit (cued recall)
test. However, subjects were only tested at 170 min post-
drug (close to oxazepam’s theoretical peak concentration)
to rule out the possible “explicit memory contamination”
explanation of the Stewart et al. implicit memory find-
ings. Consistent with previous research, both drugs im-
paired explicit memory relative to placebo. Also, both lo-
razepam and oxazepam impaired priming performance,
supporting the “time-dependence” interpretation of the
Stewart et al. findings. The results also indicate that epi-
sodic memory is impaired by both benzodiazepines in a
time-dependent fashion even when the research metho-
dology used involves everyday memory demands.
&kwd:Key words Benzodiazepine · Explicit memory ·
Implicit memory · Lorazepam · Oxazepam · Priming&bdy:
Introduction
In the last decade, memory researchers have begun to
differentiate between two main types of memory: explic-
it memory and implicit memory (e.g. Graf and Schacter
1985). Explicit memory occurs when a subject is aware
that his/her memory is being tested and consciously and
effortfully attempts remembering of previous experience.
In contrast, implicit memory is demonstrated when a
previous experience enhances an individual’s ability to
perform a task without the individual realizing that
his/her memory is being tested, or engaging in conscious
attempts at recollecting prior experience (Graf and
Schacter 1985).
Research has consistently shown that explicit memory
is impaired by a wide variety of benzodiazepines (e.g.,
alprazolam – Allen et al. 1991; lorazepam – Bishop and
Curran 1995). Early research indicated that only one
benzodiazepine, lorazepam, impaired both implicit and
explicit memory (e.g., Brown et al. 1989; Knopman
1991). Possible explanations given for this finding have
included a different cortical distribution of lorazepam, or
a different population of benzodiazepine receptors
uniquely affected by lorazepam (Curran et al. 1987;
Knopman 1991; Sellal et al. 1992; Curran and Goren-
stein 1993). In the last few years, some research has indi-
cated that lorazepam may not be the only benzodiazepine
which impairs priming (Vidailhet et al. 1994; Legrand et
al. 1995; Stewart et al. 1996). For example, Legrand and
associates found that diazepam impaired priming when
implicit memory was tested near the peak blood concen-
tration for diazepam (i.e., 60 min post-drug). This find-
ing suggests that implicit memory may have to be tested
near the specific time of peak blood concentration for the
benzodiazepine being studied, in order to observe poten-
tially detrimental effects of these drugs on implicit mem-
ory processes. Benzodiazepines are absorbed at different
speeds and these absorption rates are important predic-
tors of the clinical effects of benzodiazepines (Greenblatt
et al. 1981). For example, Curran and Gorenstein (1993)
compared lorazepam and oxazepam and found that only
lorazepam impaired priming. However, lorazepam has
been reported as having a peak absorption time of
120 min or less (see review by Greenblatt 1981), where-
as oxazepam has a peak absorption time of 162 min
S.E. Buffett-Jerrott (
✉
) · S.H. Stewart
Department of Psychology, Dalhousie University,
Life Sciences Centre, 1355 Oxford Street,
Halifax, Nova Scotia, Canada B3H 4J1
e-mail: sjerrott@is2.dal.ca
M.D. Teehan
Department of Psychiatry, Dalhousie University,
Life Sciences Centre, 1355 Oxford Street,
Halifax, Nova Scotia, Canada B3H 4J1&/fn-block:
Psychopharmacology (1998) 138:344–353 © Springer-Verlag 1998
ORIGINAL INVESTIGATION
&roles:Susan E. Buffett-Jerrott · Sherry H. Stewart
Michael D. Teehan
A further examination of the time-dependent effects of
oxazepam and lorazepam on implicit and explicit memory
&misc:Received: 3 January 1997 / Final version: 10 February 1998
(Curran et al. 1987). Therefore, Curran and Gorenstein’s
memory testing at 120 min after drug administration may
have been too early to have detected any impairments of
oxazepam on implicit memory.
Stewart et al. (1996) compared the effects of oxaze-
pam (30 mg) and lorazepam (2 mg) at two time points
(100 and 170 min post-drug) on tests of implicit and ex-
plicit memory to investigate the potential time-dependent
effects of these drugs. Stewart et al. (1996) used a cued
recall test of explicit memory and a word stem comple-
tion test of implicit memory. These tests satisfy the “re-
trieval intentionality criterion”, as they are identical ex-
cept for the instructions given to subjects (Roediger and
McDermott 1993). Therefore, the tests are directly com-
parable because any differences between responses on
the implicit and explicit tasks could only be due to the
instructions. At time 1 (100 min post-drug) the research-
ers found results identical to Curran and Gorenstein
(1993), with oxazepam impairing only explicit memory
and lorazepam impairing both implicit and explicit mem-
ory performance compared to placebo. However, at time
2 (170 min post-drug), the researchers found that both
drugs impaired priming and explicit memory.
The Stewart et al. (1996) study is consistent with the
notion of time-dependent effects of benzodiazepines on
priming. However, this latter study had several limita-
tions. The purpose of the present study was to replicate
and extend Stewart et al.’s findings after overcoming the
three main limitations of the original study.
Possibility of memory contamination
The simplest explanation for the Stewart et al. (1996)
finding is that oxazepam like other benzodiazepines (e.g.
diazepam; Legrand et al. 1995) impairs priming only
when it has reached its peak plasma concentration. How-
ever, another explanation for the Stewart et al. (1996)
findings is explicit memory contamination of the implicit
task at time 2. As noted by Schacter and associates
(1993), responses on a priming task may be affected if
subjects realize that their memory is being tested and be-
gin to use explicit remembering to complete the task.
Due to the repeated memory testing used in the Stewart
et al. (1996) study, it is possible that subjects had be-
come aware of the true purpose of the implicit task by
the second testing cycle. If this awareness caused the
subjects to complete the task using explicit memory
strategies, the implicit test can be said to have been con-
taminated by explicit memory processes, which are well
known to be impaired by benzodiazepines (Curran et al.
1995).
Lack of attentional measure
Another potential limitation of the study of Stewart et al.
(1996) is the failure to include a measure of attention.
Although the researchers did account for the effects of
the drugs on sedation, the possibility remains that the
drug-induced memory impairments were secondary to
changes in attention (Curran et al. 1987). Since benzodi-
azepines impair attention (e.g., Curran 1991), it is proba-
ble that the benzodiazepine-treated subjects in the Stew-
art et al. (1996) study would have had great difficulty at-
tending to the word list that they were trying to encode.
Lack of task involving “everyday memory” demands
Some researchers have stated that an individual’s re-
sponses on many laboratory memory tests (e.g. memory
for word lists) may not reflect his or her responses in
real-life situations (Curran 1986). The cued recall task
used by Stewart et al. (1996) may not test “everyday
memory demands” and, therefore, the observed benzodi-
azepine-induced impairments on this laboratory measure
of explicit memory may not reflect real-world memory
deficits. In fact, one study suggests that when memory
tests which more closely simulate real-life situations are
used, subjects show less benzodiazepine-induced memo-
ry impairment (Allen et al. 1991).
The present study attempted to overcome each of the
problems inherent in the Stewart et al. study as follows.
Possibility of memory contamination
The present study attempted replication of the Stewart et
al. (1996) study after omitting the first memory testing
cycle. Therefore, if oxazepam was still found to impair
priming at 170 min post-drug, explicit memory contami-
nation of the implicit task would be much less likely,
since prior exposure to the two memory tasks had been
eliminated.
Lack of attentional measure
This study included an established measure of attention
(i.e. symbol cancellation task; Mesulam 1985), which
could be used as a covariate if the subjects who received
benzodiazepines showed attentional impairments.
Lack of task involving “everyday memory” demands
A new task was added to the present study in an attempt
to assess benzodiazepine effects on “everyday memory”
abilities. For 160 min after drug administration, subjects
were shown a movie and asked questions about their
memory for movie details at successive 15-min intervals.
A memory task using a movie was included in the pres-
ent study for three reasons. First, a movie may be of
more interest to subjects than traditional memory task
stimuli such as word lists. Secondly, this novel task
served as an alternate method of testing episodic memo-
ry in addition to the cued recall task. Finally, the movie
345
memory task may involve more everyday memory de-
mands than the cued recall of a word list. Therefore, if
subjects in the benzodiazepine groups showed deficits on
the movie memory task, it would provide further evi-
dence that benzodiazepines do impair everyday memory
abilities.
The methods of the current study were identical to
those used by Stewart et al. (1996), with the exception of
the addition of the movie memory test and attention task,
and the limiting of cued word recall and word stem com-
pletion testing to 170 min post-drug. It was postulated
that both explicit and implicit memory would be im-
paired by lorazepam and oxazepam at 170 min post-
drug. This would lend support to the interpretation that
the Stewart et al. (1996) findings were not a result of ex-
plicit memory contamination of the implicit task, but in
fact reflected time-dependent effects of oxazepam in im-
pairing implicit memory processes. Based on previous
research on the relative effects of divided attention on
explicit versus implicit memory tasks (Rabinowitz et al.
1982), it was hypothesized that subjects given a benzodi-
azepine would show attentional impairments and that
these impairments would contribute to memory impair-
ments on the explicit tasks but not the priming task. On
the movie memory task, it was postulated that both the
lorazepam and oxazepam subjects would show impair-
ment on the memory questions compared to the placebo
group. It was further postulated that this impairment
would begin earlier for the lorazepam group because of
the differences in the relative absorption rates of oxaze-
pam and lorazepam (Greenblatt 1981).
Materials and methods
Subjects
Thirty healthy Dalhousie University students (15 males and 15 fe-
males) volunteered to take part in the study. Subjects ranged in age
from 19 to 36 years (mean=22.30, SD=3.44). None had a history
of alcohol or drug abuse and none were long-term users of benzo-
diazepines. Subjects were instructed to abstain from using alcohol
and other CNS drugs for 24 h before the study and to have no
more than one cup of coffee on the morning of the study. Subjects
were also allowed a light breakfast on the morning of the day of
testing. Subjects received four course credit points or $20 as com-
pensation for participating in the study. All subjects provided writ-
ten informed consent before participation. This study had ethical
committee approval.
Experimental design and drugs
Subjects were randomly assigned to one of three groups, with ten
subjects per group. The first group received 2 mg lorazepam, the
second group received 30 mg oxazepam, and the last group re-
ceived a placebo tablet. Drug tablets were given orally following a
double-blind procedure. Subjects were tested in the morning.
Tasks and procedure
Beginning 30 min prior to drug administration, subjects completed
a variety of demographic and cognitive measures to ensure pre-
drug equivalence between the three groups. First, subjects provid-
ed demographic information (age, gender, years of education).
Subjects were also weighed and had their height recorded for use
in calculation of body mass index (BMI). Given the evidence of
cross-tolerance between alcohol and benzodiazepines (see Stewart
et al. 1992), subjects were administered the Brief version of the
Michigan Alcoholism Screening Test (Brief MAST; Pokorny et al.
1972) and a measure of weekly alcohol consumption (Stewart et
al. 1995). Subjects were also administered the Drug Abuse
Screening Test (DAST; Skinner 1982), to test for drug problems.
To ensure that groups did not differ in cognitive performance
prior to drug administration, subjects completed an immediate and
delayed test of episodic memory using the Logical Memory Test
from the Wechsler Memory Scale-Revised (WMS-R; Wechsler
1987), a letter cancellation task (LCT) to assess attention (Mesu-
lam 1985), and the Vocabulary test of the Shipley Institute of Liv-
ing (SIL) Scale (Shipley 1940) to assess baseline verbal knowl-
edge. To ensure that the groups did not differ at pre-drug baseline
in their subjective levels of sedation, subjects were given five visu-
al analogue scales (VASs) in which they were asked to rate their
current state on five 100 mm lines with the following descriptors:
alert-drowsy; excited-calm; clear headed-fuzzy; energetic-lethar-
gic; and quick-slow (Danion et al. 1989). Scores on the five VASs
(number of millimeters from the left that the mark was placed)
were averaged into a mean sedation score which was used as the
dependent measure of subjective sedation. The finger tapping test
(FTT; Frith 1967) was used to evaluate baseline psychomotor
speed.
Immediately following drug administration, subjects were
shown a movie (150 min in length) which none of them had ever
seen before (i.e., Riel; Bloomfield 1979). Every 15 min the movie
was paused for 1 min and the subjects were asked to write down
the answers to five brief, orally administered open-ended ques-
tions about what had happened during that section of the movie
(e.g., “What color was Louis Riel’s jacket?”). These questions
were pilot tested with non-drugged subjects to establish a similar
difficulty level for all ten sets of questions. The pilot testing was
aimed at developing sets of questions which would enable non-
drugged subjects to achieve an average of 80% correct responses
to each of the ten sets of five movie memory questions.
The word lists used for the encoding, explicit and implicit
tasks were those used by Stewart et al. (1996) at their first post-
drug testing cycle (100 min post-drug). In the complete set of 92
words, each word had a unique three-letter word stem. The com-
plete set was divided into four sets of 23 words which were com-
bined to create two encoding sets of 46 words (Stewart et al.
1996). The four lists were balanced for number of letters, frequen-
cy of use in the English language (Thorndike and Lorge 1944),
base rate completion of the stem with the target word for a sample
of drug-free subjects (e.g. Graf and Williams 1987) and number of
possible word stem completions (Carroll et al. 1971) (see Stewart
et al. 1996).
After the movie task, subjects were again asked to complete
the five VASs to measure their post-drug sedation. They were also
given an alternate form of the attention task [symbol cancellation
task (SCT); Mesulam 1985]. These two tasks were completed be-
ginning at 160 min post-drug.
Next, subjects completed a semantic encoding task in which
they were asked to provide likability ratings to one of the two en-
coding sets of 46 words. This task took place 170 min after drug
administration. To maintain consistency with the instructional set
used by Stewart et al. (1996) in their second memory testing cycle,
the subjects were informed that they were going to be asked about
the words at a later time (intentional learning instructions). The
presentation of the two encoding lists was counter-balanced within
each drug group. The encoding task was presented on a computer
screen, with each word remaining on the screen for 5 s. Subjects
were instructed to read each word aloud and rate how much they
liked or disliked it on a 5-point scale (Brown et al. 1989; Knop-
man 1991; Curran and Gorenstein 1993; Stewart et al. 1996). This
rating was given verbally during the inter-stimulus interval (3 s).
After the subjects completed the encoding task, they were admin-
346
istered a 90-s psychomotor speed task (Symbol-Digit Test; Smith
1973) which also served as a distractor task. Next, subjects were
provided with a list of 46 word stems and asked to complete them
with the first word that came to mind (implicit task). The subjects
were not told anything about the purpose of the word-stem com-
pletion task. They were asked to complete the stems in the order
they were listed on the form and not to use proper nouns to com-
plete the stems. Subjects were given unlimited time on this task,
but were asked to work as quickly as they could. There were two
forms of the implicit task containing 23 words from the previously
shown encoding list (primed) and 23 words from the non-target
encoding list (unprimed). Presentation of the two forms of the im-
plicit task was counterbalanced within each drug group.&fnn.1:
1
Following the implicit task, subjects were given one of four
sets of the 23 stems which began the remaining words from the
encoding set which were not used for the implicit task. Subjects
were given 5 min and instructed to complete the stems with words
from the encoding task (cued recall). Subjects were instructed not
to use proper nouns to complete the word stems. Thus, this test
was identical to the implicit task, except that the instructions to
subjects differed. Subjects were told to guess if they were unsure
and were allowed to complete the stems in any order.
Results
Statistical analysis
Drug group effects were examined with separate AN-
OVAs or ANCOVAs with the aid of the Bio-Medical Da-
ta Programs (BMDP) statistical software package, Ver-
sions P4V and P2V (Dixon et al. 1985). Significant main
effects were followed by Newman-Keuls post hoc com-
parisons. Correlations between variables were computed
with the aid of the BMDP program, Version P6R. Al-
though most dependent variables were examined with
separate one-way (Drug group) ANOVAs, on the implicit
memory task, the number of stems completed with tar-
gets was analyzed with a 3×2 (Drug group × Priming
level) ANOVA with repeated measures. Likewise, for
subjective sedation, a 3×2 (Drug group × Time) ANOVA
with repeated measures was conducted which compared
subject’s pre- and post-drug VAS ratings of sedation.
Subject characteristics
There were no significant effects of Drug group for any
of the control variables. Therefore, the three groups (pla-
cebo, oxazepam and lorazepam) were equivalent in age
(mean=22.3 years, SD=3.33), gender composition
(mean=50% female) and education level (mean=3.56
years post-secondary education, SD=1.82). Groups were
also equivalent in body mass index (BMI) scores, alco-
hol problems (Brief MAST scores), drug problems
(DAST scores), and typical weekly alcohol use (drinks
per week).
Pre-drug cognitive functioning
There were no significant effects of Drug group on any
of the pre-drug cognitive measures: vocabulary, immedi-
ate and delayed episodic memory, attention, subjective
sedation or psychomotor speed.
Memory effects
Movie memory task
To ensure that the subject’s responses on the movie task
were objectively scored by the researcher who had tested
the subjects, a second individual, who had no contact
with the subjects also scored each answer. The inter-rater
reliability was very high (r=0.987). To minimize the
number of statistical comparisons on the movie memory
task data, the nine sets of movie questions (five questions
per set) were combined into three time blocks. Block 1
contained the 15 questions which asked about memory
for movie details from 16 min to 63 min post-drug ad-
ministration. Block 2 contained the 15 questions admin-
istered from 64 min to 111 min post-drug. The final
block contained the remaining 15 questions from
112 min post-drug to the completion of the movie
(160 min post-drug). The five questions which asked
about memory for the details of the first 15 min of the
movie were considered a practice phase to familiarize
subjects with the task, prior to expected onset of drug-in-
duced episodic memory impairments. By examining the
number of movie questions that each drug group correct-
ly answered at each time block (Fig. 1), a distinct pattern
can be observed. It appears that the placebo group
showed relatively constant performance across the trial
blocks. In contrast, the lorazepam group shows rapid de-
terioration of performance across time blocks, with the
performance of the oxazepam group falling mid-way be-
tween that of the other two groups by the third time
block.
Movie memory performance for each Time block was
examined with relation to the initial hypotheses by de-
composing the full 3×3 (Drug group × Time block) table
of means in a series of planned comparisons (Tabachnick
and Fidell 1989). First, we postulated that the lorazepam
group would show impaired memory compared to the
placebo group by Time block 2 of the movie memory
test and that this impairment would also be evident at
Time block 3. These hypotheses were evaluated by com-
paring the placebo and lorazepam group at each of the
three Time blocks. At Time block 1, there was no signifi-
cant difference between the two groups [F(1,18)=2.25,
NS]. At Time block 2, there was a marginally significant
difference between the two groups, with the lorazepam
group showing poorer performance [F(1,18)=3.01,
P<0.10]. At Time block 3, the lorazepam group was sig-
nificantly impaired relative to the placebo group
[F(1,18)=12.73, P<0.01]. We also predicted that the ox-
azepam group would show impaired memory compared
347
1
Counter-balancing for content of the implicit test across
primed/unprimed conditions was achieved through the use of two
encoding sets and two forms of the implicit task. For example, for
form 1 of the implicit task the same stems that served as “primed”
stems for a subject who received encoding set 1, would serve as
“unprimed” stems for a subject who received encoding set 2
348
to the placebo group, but that this impairment would not
be evident until Time block 3. To evaluate this hypothe-
sis, the performances of the placebo and oxazepam sub-
jects on the movie memory questions were compared at
each Time block. At Time block 1 and Time block 2,
there were no significant differences between the two
groups [F(1,18)=1.40, NS; F(1,18)=1.12, NS, respective-
ly]. However, at Time block 3, the oxazepam group was
impaired relative to the placebo group [F(1,18)=6.11,
P<0.05]. Overall, these results supported our hypotheses
that both drug groups would show impaired memory per-
formance when compared to the placebo group, and that
the lorazepam group would show impaired memory per-
formance at an earlier time point than the oxazepam
group due to lorazepam’s faster rate of absorption
(Greenblatt 1981).
Explicit memory
Cued recall performance was scored as the total number
of word stems correctly completed with target words
from the semantic encoding task. Means (and SDs) for
the cued recall task, as a function of Drug group, are dis-
played in Fig. 2. The analysis revealed a significant main
effect of Drug group [F(2,27)=3.51, P<0.05]. Further an-
alyses revealed a significant difference between the pla-
cebo and lorazepam groups [Q(3,27)=4.97, P<0.01) and
between the placebo and oxazepam groups [Q(2,27)=
4.06, P<0.01], with the placebo group correctly complet-
ing more word stems in each case. There was no signifi-
cant difference between the performance of the loraze-
pam and oxazepam groups [Q(2,27)=0.92, NS] on this
task (see Fig. 2).
Implicit memory
Analysis of the total number of word stems completed
with any word by each drug group was not significant
[F(2,27)=0.17, NS]. No subject completed fewer than 41
of the 46 word stems. The implicit memory test was
scored as the number of primed word stems and un-
primed word stems which were completed with target
words from the encoding lists. Means (and SDs) for the
number of stems completed with targets (as a function of
Drug group and Priming level) are shown in Fig. 3. Re-
sults indicate a main effect of Drug group [F(2,27)=4.51,
P<0.05], a main effect of Priming level [F(1,27)=59.94,
P<0.0001] and a Drug group × Priming level interaction
[F(2,27)=4.07, P<0.05]. The main effect of Priming lev-
el was due to a greater number of primed word stems be-
Fig. 1 Number of movie questions correctly answered at three
post-drug time blocks as a function of drug group. Bars represent
standard errors. ●● Placebo (n=10), ▲ oxazepam (n=10), ■■ loraze-
pam (n=10)&/fig.c:
Fig. 2 Explicit memory performance: mean number of words cor-
rectly recalled on the cued recall test, as a function of drug group.
Bars represent standard deviations&/fig.c:
349
ing completed with targets compared to unprimed stems.
The main effect of Drug group was due to the placebo
group completing more word stems with targets than the
active drug groups. To investigate the interaction further,
simple effects of Drug group were examined at each lev-
el of Priming. For the unprimed word stems, there was
no simple main effect of Drug group [F(2,27)=0.97, NS].
Therefore, drug administration did not affect subject’s
chance performance in generating unprimed targets.
However, for the primed stems, there was a significant
simple main effect of Drug group [F(2,27)=5.44,
P<0.05]. Further analysis indicated that the placebo
group generated significantly more primed targets than
both the lorazepam [Q(3,27)=6.51, P<0.01] and oxaze-
pam groups [Q(2,27)=4.16, P<0.01]. The performance of
the oxazepam and lorazepam groups did not differ sig-
nificantly [Q(2,27)=2.35, NS]. Completion of primed
stems with targets was about three-and-a-quarter times
chance levels for placebo, but only two to two-and-a-half
times chance for the lorazepam and oxazepam groups
(see Fig. 3).
Subjective sedation
Results indicated no significant main effect of Drug
group [F(2,27)=0.57, NS]. However, there was a signifi-
cant main effect of Time [F(2,27)=23.31, P<0.0001] and
a significant Drug group × Time interaction [F(2,27)=
9.34, P<0.001]. To explore this interaction further, an
analysis of simple effects of Drug group was performed
at each time. There was no significant simple main effect
of Drug group before drug administration [F(2,27)=2.22,
NS]. However, after drug administration, there was a sig-
nificant main effect of Drug group [F(2,27)=6.49,
P<0.01]. Mean (and SD) post-drug subjective sedation
ratings are illustrated in Table 1a as a function of Drug
group. Further analysis indicated that the placebo group
was less sedated after drug administration than the lora-
zepam group [Q(3,27)=7.06, P<0.01] and the oxazepam
group [Q(2,27)=4.78, P<0.01] (see Table 1a). The oxaze-
pam group and lorazepam group were not significantly
different from each other [Q(2,27)=2.28, NS] (see Table
1a). Simple effects of Time were also performed for each
Drug group. There was a significant simple main effect
of Time for the lorazepam [F(1,27)=29.16, P<0.0001]
and oxazepam [F(1,27)=12.49, P<0.01] groups, with
greater subjective sedation ratings post-drug compared to
pre-drug. However, there was no effect of Time for the
placebo group [F(1,27)=0.33, NS].
Psychomotor speed
Means (and SDs) for the symbol digit task are shown in
Table 1b, as a function of Drug group. The number of
items correctly completed in 90 s on this psychomotor
speed task by the three groups showed a similar pattern
to those obtained by Stewart et al. (1996), with the ben-
zodiazepine subjects appearing to be somewhat impaired
compared to the placebo subjects (see Table 1b). Howev-
er, the Drug group main effect did not prove statistically
significant [F(2,27)=1.77, NS].
Attention task
Mean number of omissions on the symbol cancellation
test (SCT) (Mesulam 1985) and time taken to complete
the SCT were scored as post-drug measures of attention
(Curran et al. 1987). Means (and SDs) for these atten-
tional measures are shown in Table 1c as functions of
Drug group. There was no significant effect of Drug
group [F(2,27)=0.53, NS] on SCT omissions (see Table
1c). Analysis of the time taken to complete the SCT re-
vealed a trend for a Drug group effect [F(2,27)=2.55,
P=0.10] (see Table 1c). As Bishop and Curran (1995)
have reported significant lorazepam-induced impair-
ments on the same attention task, post-hoc comparisons
were performed. The placebo group completed the SCT
faster than the lorazepam [Q(3,27)=4.49, P<0.05] group
(see Table 1c). However, the oxazepam group did not
differ significantly from the placebo [Q(2,27)=2.64, NS]
or lorazepam [Q(2,27)=1.85, NS] groups (see Table 1c).
Covariation between cognitive impairments
and performance on the memory tasks
A series of analyses of covariance (ANCOVAs) were
conducted which covaried out post-drug subjective seda-
Fig. 3 Implicit memory performance: mean number of stems
completed with target on the word-stem completion test, as a func-
tion of drug group and priming. Bars represent standard devia-
tions. ■ Unprimed, ■■ primed
350
tion and attention time, respectively, from the scores on
the implicit task (word-stem completion), explicit task
(cued-recall), and the last time block of the movie mem-
ory task. Covarying subjective sedation and attention
time scores did not affect the pattern of Drug group dif-
ferences or the significance levels of Drug group effects
on the implicit memory task. Also, covarying subjective
sedation did not affect the pattern of Drug group differ-
ences or the significance levels of the Drug group effects
on the cued recall task. However, for the cued recall task,
treating attention time as a covariate did not change the
pattern of drug group means, but did reduce the Drug
group effect to marginal significance [F(2,26)=3.01,
P=0.07]. Although covarying subjective sedation did not
affect the drug group differences on the cued recall and
word-stem completion tasks, it did influence the drug
group differences on the last time block of the movie
memory task: the lorazepam group was still significantly
impaired relative to placebo [F(1,17)=6.44, P<0.05] but
the difference between the oxazepam and placebo group
was reduced to marginal significance [F(1,17)=3.16,
P=0.09]. Covarying attention time scores affected the re-
sults of the last time block of the movie memory task in
a similar fashion. The lorazepam-treated subjects contin-
ued to evidence significant movie memory impairments
relative to placebo [F(1,17)=13.28, P<0.01], but the low-
er scores of the oxazepam group relative to placebo were
no longer significantly different [F(1,17)=2.64, NS].
Relation between stem completion and cued recall
A priming index, calculated by subtracting the number
of unprimed stems completed with target words from the
number of primed target completions, was determined
for each subject. Subjects’ priming index scores were
correlated with the number of correctly completed word
stems on the explicit task. The resulting correlation coef-
ficient was positive but not significant (r=0.204, NS), in-
dicating that performance on the implicit task was rela-
tively independent of explicit memory task performance.
Discussion
Overall, this study lends stronger support to the hypothe-
sis proposed in the Stewart et al. (1996) study, that ben-
zodiazepines have a time-dependent effect on implicit
memory. The present results also support previous find-
ings of benzodiazepine’s impairments of explicit memo-
ry (e.g. Danion et al. 1990). In addition, this study ex-
tends previous findings of benzodiazepine-induced mem-
ory impairments to another explicit memory measure
which may have more interest to subjects (i.e., the movie
memory task). Finally, this study is an important addition
to previous research which indicates that there is a disso-
ciation between the sedative and cognitive effects of ben-
zodiazepines (e.g., Hommer et al. 1993).
Even though the three groups were initially equivalent
on all pre-drug cognitive measures, both oxazepam and
lorazepam impaired priming compared to the placebo
group at 170 min post-drug. Unlike the study of Stewart
et al. (1996), there was no increased potential for explicit
contamination of the implicit task due to repeated testing
with the same memory tasks, as the subjects in the pres-
ent study were only tested with the cued recall and word-
stem completion tasks at one time point. Therefore, it ap-
pears that the most likely explanation for this finding is
that when oxazepam is nearing its peak blood concentra-
tion (Greenblatt 1981), it begins to exert impairments on
implicit memory processes.
The present study used memory tasks that satisfied
the retrieval intentionality criterion and, therefore, are di-
rectly comparable (Roediger and McDermott 1993). Al-
though a dissociation between explicit versus implicit
memory was not observed directly as a function of drug
Table 1a–c Means (and SDs)
on the post-drug measures of
sedation and attention as
functions of drug group. a Sub-
jective sedation, b Psychomo-
tor/cognitive speed, c attention.
Subjective sedation = mean of
five 0–100 visual analogue
scales (VASs) of subjective se-
dation (Danion et al. 1989);
Psychomotor/cognitive speed =
post-drug symbol digit test
(Smith 1973) score. Attention =
number of omissions on and
time taken to complete symbol
cancellation task (SCT; Mesu-
lam 1985). All comparisons are
across drug groups&/tbl.c:&tbl.b:
Measure Drug group
Placebo Oxazepam Lorazepam
(n=10) (n=10) (n=10)
a Subjective sedation
Mean VAS rating 49.14
ab
65.74
b
73.64
a
(4.74) (5.14) (4.84)
b Psychomotor speed
Mean Symbol Digit Test score 62.10 57.80 54.60
(9.35) (9.29) (8.17)
c Attentional measures
Attentional ability I
Number of omissions on SCT 1.10 0.50 0.70
(1.85) (1.08) (0.82)
Attentional ability II
Time to complete SCT (sec) 61.20
c
72.20 79.90
c
(11.21) (15.27) (26.10)
Means with similar super-
scripts are significantly differ-
ent from one another:
a
P<0.01,
b
P<0.01,
c
P<0.05&/tbl.bP
administration in the present study, the fact that there
was no significant correlation between subjects’ perfor-
mance on the word-stem completion task and cued recall
task lends support to the notion of two separate memory
systems (Schacter 1995). If a subject’s ability to perform
on the implicit task is not related to his or her perfor-
mance on the directly comparable explicit task, it is
probable that the two memory tasks are actually measur-
ing different constructs.
Many researchers have noted that traditional laborato-
ry tests of memory may lack “ecological validity” (i.e.,
are not representative of real-life memory requirements)
(e.g., Curran 1986). When subjects in the present study
were administered a task which appears to be more rep-
resentative of real-life memory requirements (i.e., the
novel movie memory task), benzodiazepines were still
found to impair explicit memory performance. Both the
oxazepam and lorazepam groups showed impairments
relative to placebo in their ability to recall details per-
taining to the story-line of a movie from 112 min to
160 min post-drug. Compared to the oxazepam group,
the lorazepam-treated subjects showed a pattern of earli-
er episodic memory decline, with impairments marginal-
ly evident relative to placebo at the 64–111 min post-
drug testing point. These relative rates of episodic mem-
ory impairments across the two drugs compared to place-
bo are consistent with their relative absorption speeds
(Greenblatt 1981). However, empirical data on the actual
ecological validity of this novel task need to be gathered
before any definitive conclusions about benzodiazepines’
effects on real-world memory abilities can be made.
In addition to impairing memory, benzodiazepines
have many other effects on cognitive processes including
effects on sedation and attention (Curran 1986). In the
present study, the drugged subjects reported more seda-
tion than placebo subjects. To help rule out the possibili-
ty of subjective sedation causing the memory impair-
ments observed in the present study, a series of analyses
of covariance were performed. The subject’s reported se-
dation did not influence drug group effects on priming
on the implicit memory (word-stem completion) task or
the cued recall test. This is consistent with previous re-
search which indicated that the cognitive effects of ben-
zodiazepines do not appear to be dependent on their sed-
ative effects (e.g., Hommer et al. 1993). However, sub-
jective sedation appeared to contribute partially to ben-
zodiazepine-induced impairments in movie memory per-
formance. It is thus possible that subjective sedation may
have a greater effect on explicit memory tasks that simu-
late “real-life” situations as compared to traditional labo-
ratory tasks.
In the present study, drugged subjects were also mar-
ginally slower on an attention test (speed of completion
of the symbol-cancellation task), than the placebo sub-
jects. Since Stewart and associates (1996) failed to in-
clude an attention task, it is possible that their drugged
subjects were attending less to the word lists at encoding
than placebo subjects, and that the memory impairments
observed were secondary to these attentional impair-
ments. To help rule out this possibility in the present
study, a series of analyses of covariance were performed.
As with sedation, subject’s attentional impairments did
not influence the drug group effects of priming on the
implicit memory (word-stem completion) task. However,
attention did appear to contribute to the drug group dif-
ferences in explicit memory performance [i.e., cued re-
call (cf. Bishop and Curran 1995), as well as movie
memory, performance]. These findings are consistent
with the results of previous research, which found that
level of processing and divided attention manipulations
influence explicit memory performance but do not affect
priming (Rabinowitz et al. 1982). Overall, the fact that
attentional impairments did not contribute to drug-in-
duced impairments of implicit memory performance but
did contribute to explicit memory performance lends fur-
ther support to the idea that implicit and explicit memory
processes are dissociable (Roediger and McDermott
1993).
In the future, the effects of other benzodiazepines on
implicit memory should be tested close to the time of
peak blood concentration for the benzodiazepine in ques-
tion. To date, only diazepam (Legrand et al. 1995), lora-
zepam (e.g., Curran and Gorenstein 1993; Bishop et al.
1996) and oxazepam (Stewart et al. 1996) have been
shown to impair priming; the results of each of these lat-
ter studies are consistent with the time-dependence hy-
pothesis. In future studies, benzodiazepines should also
be tested at various time points to gain a better under-
standing of the impairment curves for both implicit and
explicit memory. It is recommended that such future
studies be conducted as between-subjects designs to re-
duce the possibility of repeated testing contaminating the
implicit task (Roediger and McDermott 1993). Based on
previous research (Stewart et al. 1996), it appears that
oxazepam-induced explicit memory impairments be-
come apparent before impairments of implicit memory.
Specifically, it appears that oxazepam impairs explicit
memory, but not implicit memory, if memory is tested
well before the time of peak blood concentration. In con-
trast, near the time of peak plasma concentrations of ox-
azepam, this drug impairs both explicit and implicit
memory. Future research will need to examine the effects
of oxazepam on implicit and explicit memory following
the time of its peak blood concentration to determine the
persistence of both implicit and explicit memory impair-
ments. Also, as dose-dependent effects of benzodiaze-
pines on explicit memory have been reported (see Wein-
gartner et al. 1995), future research should examine dif-
ferent doses of oxazepam and lorazepam to determine
how dosage is related to the impairment of implicit
memory.
It should be noted that explicit and implicit memory
have been tested in lorazepam-treated subjects both
slightly before, near, and well after, the theoretical peak
plasma concentration of lorazepam of approximately
120 min post-drug. Thus far, no dissipation of loraze-
pam-induced explicit memory or implicit memory im-
pairments relative to placebo have been noted (i.e., Stew-
351
art et al. (1996) 100 and 170 min post-drug; Curran and
Gorenstein (1993) 120 min post-drug; present study
170 min post-drug). Lorazepam has thus been shown to
produce impairments on both explicit and implicit mem-
ory tasks relative to placebo from 100 to 170 min post-
drug. Although these findings do not directly support the
time-dependence interpretation, it is possible that the
memory testing time may have to be extended more than
50 min post-peak, and reduced to more than 20 min pre-
peak, to observe the postulated dissociation between im-
plicit and explicit memory impairments. In fact, for lora-
zepam, Greenblatt and associates (1981) have stated that
the duration of an acute dose of lorazepam tends to be
extended because the distribution of lorazepam is less
extensive than the distribution of other benzodiazepines.
Therefore, high concentrations of lorazepam can remain
in the plasma for many hours.
It is important to note that implicit memory is no lon-
ger viewed as a single entity by memory researchers (e.g.
Tulving and Schacter 1990). One of the several distinc-
tions which has been made recently in the literature is
between perceptual and conceptual priming, which pre-
liminary evidence suggests are distinct constructs (e.g.,
Keane et al. 1991). To add support to this hypothesis, re-
search needs to determine if perceptual and conceptual
priming can be dissociated by benzodiazepines. Since re-
searchers do not yet agree on the type of priming that is
being measured by the word-stem completion task (e.g.,
Keane et al. 1991 versus Roediger and McDermott
1993), future research should examine the time depen-
dent effects of various benzodiazepines on implicit mem-
ory tasks that are established as measures of conceptual
versus perceptual priming.
Overall, it appears that benzodiazepines cause impair-
ments of both explicit and implicit memory if they are
tested near the time of peak blood concentrations for the
specific drug in question. Benzodiazepine-induced ex-
plicit memory impairments are apparent even when the
memory task administered appears to more closely simu-
late real-life memory requirements. These demonstrated
impairments of both implicit and explicit memory by
benzodiazepines may make it difficult to effectively use
concurrent cognitive or behavioral therapy in the treat-
ment of individuals with anxiety disorders who also take
benzodiazepines (Wardle 1990; Curran 1991). Although
chronic benzodiazepine users become tolerant to the sed-
ative and attentional effects of benzodiazepines, they
continue to suffer from memory impairments and it is
likely that these difficulties have severe implications for
their everyday functioning (Curran 1992).
&p.2:Acknowledgements S.E.B.-J. is supported by a post-graduate
scholarship from the National Sciences and Engineering Research
Council of Canada (NSERC). This study was supported by an op-
erating grant from NSERC and a Dalhousie Research Develop-
ment Funds for Sciences (RDFS) grant awarded to S.H.S.
References
Allen D, Curran HV, Lader M (1991) The effects of repeated dos-
es of clomipramine and alprazolam on physiological, psycho-
motor and cognitive functions in normal subjects. Eur J Clin
Pharmacol 40: 355–362
Bishop KI, Curran HV (1995) Psychopharmacological analysis of
implicit and explicit memory: a study with lorazepam and the
benzodiazepine antagonist flumazenil. Psychopharmacology
121: 267–278
Bishop KI, Curran HV, Lader M (1996) Do scopolamine and lora-
zepam have dissociable effects on human memory systems? A
dose-response study with normal volunteers. Exp Clin Psycho-
pharmacol 4: 292–299
Bloomfield G (Director), Colbert S (Executive Producer), Trent J
(Producer) (1979) Riel (Film). CBC and Green River Pictures
Inc.
Brown MW, Brown J, Bowes JB (1989) Absence of priming cou-
pled with substantially preserved recognition in lorazepam-in-
duced amnesia. Q J Exp Psychol 41A: 599–617
Carroll JB, Davies P, Richman B (1971) Word frequency book.
Houghton-Mifflin, Boston
Curran HV (1986) Tranquilizing memories: a review of the effects
of benzodiazepines on human memory. Biol Psychol 23:
179–213
Curran HV (1991) Benzodiazepines, memory and mood: a review.
Psychopharmacology 105: 1–8
Curran HV (1992) Memory functions, alertness and mood of long-
term benzodiazepine users: a preliminary investigation of the
effects of a normal daily dose. J Psychopharmacol 6: 69–75
Curran HV, Gorenstein C (1993) Differential effects of lorazepam
and oxazepam on priming. Int Clin Psychopharmacol 8: 37–42
Curran HV, Schiwy W, Lader M (1987) Differential amnesic prop-
erties of benzodiazepines: a dose-response comparison of two
drugs with similar plasma elimination half-lives. Psychophar-
macology 92: 358–364
Curran HV, Barrow S, Weingartner H, Lader M, Bernik M (1995)
Encoding, remembering and awareness in lorazepam-induced
amnesia. Psychopharmacology 122: 188–193
Danion JM, Zimmerman MA, Willard-Schroeder D, Grange D,
Singer L (1989) Diazepam induces a dissociation between ex-
plicit and implicit memory. Psychopharmacology 99: 238–243
Danion JM, Zimmerman MA, Willard-Schroeder D, Grange D,
Welsch M, Imbs JL, Singer L (1990) Effects of scopolamine,
trimipramine and diazepam on explicit memory and repetition
priming in healthy volunteers. Psychopharmacology 102:
422–424
Dixon W, Brown MB, Engelman L, Frane JW, Hill MA, Jennrich
RI, Toporek JD (eds) (1985) BMDP statistical software manu-
al. University of California Press, Berkely, Calif.
Frith CD (1967) The effects of nicotine on tapping: II. Life Sci 6:
321–326
Graf P, Schacter DL (1985) Implicit and explicit memory for new
associations in normal and amnesic subjects. J Exp Psychol
11: 501–518
Graf P, Williams D (1987) Completion norms for 40 three-letter
word-stems. Behav Res Meth Instr Comput 19: 422–445
Greenblatt DJ (1981) Clinical pharmacokinetics of oxazepam and
lorazepam. Clin Pharmacokinet 6: 89–105
Greenblatt DJ, Shader RI, Divoll M, Harmatz JS (1981) Benzodi-
azepines: a summary of pharmacokinetic properties. Br J Clin
Pharmacol 11: 11S–16S
Hommer D, Weingartner HJ, Brier A (1993) Dissociation of ben-
zodiazepine induced amnesia from sedation. Psychopharma-
cology 112: 455–460
Keane MM, Gabrieli DE, Fennema AC, Growdon JH, Corkin S
(1991) Evidence for a dissociation between perceptual and
conceptual priming in Alzheimer’s disease. Behav Neurosci
105: 326–342
Knopman D (1991) Unaware learning versus preserved learning in
pharmacologic amnesia: similarities and differences. J Exp
Psychol [Learn Mem Cognit] 17: 1017–1029
352
Stewart SH, Pihl RO, Padjen AL (1992) Chronic use of alcohol
and/or benzodiazepines may account for evidence of altered
benzodiazepine receptor sensitivity in panic disorder. Arch
Gen Psychiatry 49: 329–330
Stewart SH, Peterson JB, Pihl JB (1995) Anxiety sensitivity and
self-reported alcohol consumption rates in university women. J
Anx Disord 9: 283–292
Stewart SH, Rioux GF, Connolly JF, Dunphy SC, Teehan MD
(1996) Effects of oxazepam and lorazepam on implicit and ex-
plicit memory: evidence for possible influences of time
course. Psychopharmacology 128: 139–149
Tabachnick BG, Fidell LS (1989) Using multivariate statistics
(2nd edn). Harper Collins, New York
Thorndike EL, Lorge I (1944) The teacher’s word book of 30,000
words. Columbia University Teachers College Press, New
York
Tulving E, Schacter DL (1990) Priming and human memory sys-
tems. Science 247: 301–306
Vidailhet P, Danion JM, Kauffmann-Muller F, Grange D, Giersch
A, Van Der Linden M, Imbs JL (1994) Lorazepam and diaz-
epam effects on memory acquisition in priming tasks. Psycho-
pharmacology 115: 397–406
Wardle J (1990) Behavior therapy and benzodiazepines: allies or
antagonists? Br J Psychiatry 156: 163–168
Wechsler D (1987) Manual for the Wechsler Memory Scale – re-
vised (WMS-R). Psychological Corporation, New York
Weingartner HJ, Sirocco K, Curran V, Wolkowitz O (1995) Mem-
ory facilitation following the administration of the benzodiaze-
pine triazolam. Exp Clin Psychopharmacol 3: 298–303
353
Legrand F, Vidailhet P, Danion JM, Grange D, Giersch A, Van Der
Linden M, Imbs JL (1995) Time course of the effects of diaz-
epam and lorazepam on perceptual priming and explicit mem-
ory. Psychopharmacology 118: 475–479
Mesulam MM (1985) Verbal and nonverbal cancellation tests.
Principles of behavioral neurology. F.A. Davis, Philadelphia,
pp 102–103
Pokorny AD, Miller BA, Kaplan HB (1972) The Brief MAST: a
shortened version of the Michigan alcoholism screening test.
Am J Psychiatry 129: 342–345
Rabinowitz JC, Craik FI, Ackerman BP (1982) A processing re-
source account of age differences in recall. Can J Psychol 36:
325–344
Roediger HL III, McDermott KB (1993) Implicit memory in nor-
mal human subjects. In: Spinnler H, Boller F (eds) Handbook
of neuropsychology (vol 9). Elsevier, Amsterdam, pp 63–131
Schacter DL (1995) Implicit memory: a new frontier for cognitive
neuroscience. In: Gazzaniga MS (ed) The cognitive neurosci-
ences. M.I.T. Press, Cambridge, Mass., pp 815–824
Schacter DL, Chiu CY, Ochsner KN (1993) Implicit memory: a
selective review. Annu Rev Neurosci 16: 159–182
Sellal F, Danion JM, Kauffmann-Muller F, Grange D, Imbs JL,
Van Der Linden M, Singer L (1992) Differential effects of di-
azepam and lorazepam on repetition priming in healthy volun-
teers. Psychopharmacology 108: 371–379
Shipley WC (1940) A self-administering scale for measuring intel-
lectual impairment and deterioration. J Psychol 9: 371–377
Skinner HA (1982) The drug abuse screening test. Addict Behav
7: 363–371
Smith A (1973) Symbol digit modalities test. Western Psychologi-
cal Services, Los Angeles, CA