Memory encoding and retrieval on the ascending and descending

Psychopharmacology (2005) 182: 305–317

DOI 10.1007/s00213-005-0096-2

ORIGINAL INVESTIGATION

Hedvig Söderlund

Elizabeth S. Parker

Barbara L. Schwartz

Endel Tulving

Memory encoding and retrieval on the ascending and descending

limbs of the blood alcohol concentration curve

Received: 8 June 2005 / Accepted: 10 June 2005 / Published online: 14 September 2005

# Springer-Verlag 2005

Abstract Rationale: Little is known about acute effects

of alcohol on memory encoding and retrieval on different

limbs (ascending and descending) of the blood alcohol

concentration (BAC) curve. Objectives: This extensive

experiment was designed to examine alcohol’s effects on

memory encoding and retrieval throughout a protracted

drinking episode. Methods: In a 9-h session, male partic-

ipants consumed either alcohol (1 ml/kg) or placebo (n=

32/32) over a period of 90 min and learned various ma-

terials in different memory tasks before, during, and after

consuming the drinks, while their BAC levels were mon-

itored. A week later, in a similar session, they were tested

on learned materials before, during, and after drinking.

Mood was assessed throughout both sessions. Results: Al-

cohol impaired recall of words more than recognition, and

cued recall most severely. Perceptual priming and picture

recognition were not affected by alcohol. Alcohol impaired

encoding in cued recall, recognition of completed word

fragments, and free recall regardless of limb, but impaired

retrieval in word recognition only during the ascending

BAC. Alcohol increased negative mood on the descending

limb during the first session, and on the ascending limb

during the second session. Conclusions: Under naturalis-

tic drinking conditions, alcohol’s effects on memory depend

on task, memory process, and limb of the BAC curve. The

differential effects of alcohol on retrieval during the as-

cending and descending limbs demonstrate the importance

of examining the differential effects on the two limbs.

Keywords Alcohol

Memory

Encoding

Retrieval

Mood

Ascending

Descending

Limb

Introduction

Psychopharmacological research on alcohol’s acute effects

on memory is relevant for understanding the cognitive

impairments produced by one of the most widely enjoyed

and abused drugs in western cultures. Research on alcohol

is also relevant to basic questions about human memory

function. Cognitive psychology has strongly influenced

research on alcohol and human memory, and alcohol has

been found to impair performance in certain memory

tasks while leaving that in others unaffected. Episodic

memory (Tulving 1972, 1983) is particularly impaired by

alcohol (Hashtroudi et al. 1984; Nilsson et al. 1989;

Curran and Hildebrandt 1999), semantic retrieval is some-

what less so (Hartocollis and Johnson 1956; Wendt and

Risberg 2001), whereas priming (i.e., “the facilitated iden-

tification of perceptual objects from reduced cues as a

consequence of a specific prior exposure to an object”;

Schacter 1994) is regularly unaffected (Fillmore et al. 2000;

Hashtroudi et al. 1984; Nilsson et al. 1989; Duka et al.

2001).

H. Söderlund (*)

The Rotman Research Institute,

Baycrest Centre for Geriatric Care,

3560 Bathurst Street,

Toronto, ON, M6A 2E1, Canada

e-mail: [email protected]

Tel.: +1-416-7852500

Fax: +1-416-7852862

E. S. Parker

Department of Neurology,

University of California,

Irvine, CA, USA

Department of Psychiatry and Neurology,

University of Southern California,

Los Angeles, CA, USA

B. L. Schwartz

Department of Psychiatry,

Washington Veterans Affairs Medical Center,

Washington, DC, USA

Georgetown University School of Medicine,

Washington, DC, USA

E. Tulving

The Rotman Research Institute,

Baycrest Centre for Geriatric Care,

Toronto, Canada

Department of Psychology,

University of Toronto,

Toronto, Canada

Although sophisticated tasks have been used to assess

alcohol’s effect on different aspects of memory, the phar-

macological and metabolic characteristics of alcohol have

received less attention. The procedure for administering

alcohol is crucial, because alcohol has complex effects on

neurotransmission in different brain regions depending on

variations in dose, rate and route of administration, and

other variables that can produce neural excitation, depres-

sion, or both (Eckardt et al. 1998). The conclusions drawn

from previous research are mainly based on paradigms

where a bolus dose of alcohol has been consumed in less

than 30 min (but see Ilan and Gevins 2001) and where only

one or two forms of memory were assessed at one point

of the blood alcohol concentration (BAC) curve, usually

at peak. These experiments have thrown little light on the

effects of alcohol at different points on the ascending and

descending limbs of the BAC curve, although differential

effects of the two limbs on cognition, mood, and phys-

iological responses have been published. Memory, ab-

stract reasoning, attention, and reaction/anticipation time

have been found to be impaired on the ascending limb

(Hurst and Bagley 1972; Jones and Vega 1972; Jones

1973; Nicholson et al. 1992), a limb usually associated

with an increase in skin conductance (Pishkin et al. 1983),

arousal and positive mood, but also with aggression. The

descending limb usually produces sedation, negative mood

(Babor et al. 1983; Sutker et al. 1983; Lukas et al. 1986;

Giancola and Zeichner 1997; Papineau et al. 1998; Erblich

and Earleywine 2003), and impaired executive function

(Pihl et al. 2003). However, the biphasic effects of alcohol

on mood, psychomotor/cognitive performance, and heart

rate can vary between individuals depending on their

family history and drinking habits (Conrod et al. 1997;

Hiltunen 1997; Holdstock and de Wit 1998; King et al.

2002).

Alcohol usually impairs episodic memory encoding

more than episodic memory retrieval (e.g., Goodwin et al.

1969; Birnbaum et al. 1978; Petersen 1977; but see

Fillmore et al. 1999), although few studies have made

explicit comparisons between the two processes. Encoding

and retrieval often take place during the same drinking

session, making it difficult to draw conclusions about

alcohol’s differential effect on the two processes.

The purpose of the present study was to examine

alcohol’s effect on memory encoding and retrieval with a 1-

week retention interval. In addition, alcohol was adminis-

tered gradually, as it might be consumed in social drinking

situations. Subjects were assessed throughout the whole

BAC curve using four different memory tasks. During

day 1, or the “study session”, subjects studied different

materials at eight different times across a 9-h session. On

day 8, a week later, subjects were tested for the studied

materials during a comparable session, the “test session”.

By combining data from the study and test sessions, both

encoding and retrieval were assessed while subjects were

sober, on the ascending limb, at peak, and on the de-

scending limb of the BAC curve. A control group was

included that followed the exact same procedure, only

without alcohol, to control for sequential effects, practice,

and fatigue. Mood was assessed across the BAC curve

during both sessions.

We expected that alcohol would differentially affect the

memory tasks, impairing episodic memory, but not prim-

ing. Within episodic memory, performance in free recall

should be the most impaired, given that this task offers the

fewest cues at retrieval. Further, as previously shown using

other memory tasks, alcohol should impair encoding more

than retrieval and lead to worse performance on the

ascending limb than on the descending limb at equivalent

BACs.

Methods

Subjects

Sixty-four men between 21 and 29 years of age (M, 22.8;

SD, 2.4) were recruited from universities in the Washington

DC area and paid to participate in the study. Only men were

included to eliminate the risk of giving alcohol to pregnant

women. Prior to inclusion, subjects were screened for

drinking habits, and a psychiatrist screened them to exclude

persons with medical and/or psychiatric problems. Subjects

reported consuming an average of 3.9 (±2.1) drinks, 2.3

(±1.4) times per week. Drinking frequency ranged from

once a month (one person) to once a day (one person), with

a median of once to twice a week. Written informed

consent was obtained in accordance with the IRB approved

protocol.

Subjects were randomly assigned to either the alcohol or

placebo group, with 32 persons in each. The two groups

did not differ in terms of age, alcohol consumption, or

intelligence score (Shipley 1940; cf. Zachary 1986). The

placebo group was slightly more educated than the alcohol

group [F

(1,62)

=3.9, p=0.05], but the two groups had equiva-

lent sober memory performance.

Design

On day 1, subjects studied one eighth of the materials from

each task during each of eight successive study periods

(SPs), initiated at the same time points for all subjects, and

lasting approximately 25 min. In the alcohol group, these

corresponded to (see Fig. 1a): (1) 60 min before the first

drink (−60 min), (2) 25 min before the first drink (−25 min),

(3) low-ascending BAC (∼0.03 g/100 ml; +10 min), (4)

high-ascending BAC (∼0.06 g/100 ml; +60 min), (5) peak

BAC (∼0.08 g/100 ml; +120 min), (6) high-descending

BAC (∼0.06 g/100 ml; +200 min), (7) low-descending

BAC (∼0.03 g/100 ml; +315 min), and (8) zero BAC

(<0.01 g/100 ml; +440 min).

A week later, subjects had a second identical drinking

episode and were tested on the materials they had studied

the week before. Because test procedures were longer than

study procedures, the tasks could not be administered as

many times at test (day 8) as they were at study (day 1).

Accordingly, all four tasks were performed while subjects

306

were sober (− 60 min), only once on the ascending limb

(either at low or high BAC; +10 or +60 min), at peak BAC

(+110 min), and once on the descending limb (either at low

or high BAC; +200 min or +315 min; see Fig. 1b). Like the

study periods on day 1, test periods (TPs) occurred at the

same time points for all subjects.

Using this approach, the eight study periods were

orthogonally combined with four test periods to yield 7-

day retention data for each of the four tasks, resulting in 32

study/test combinations. The 32 combinations systemat-

ically varied within each subject with respect to the subject’s

BAC at study (sober, ascending limb, peak, descending

limb, and sober postintoxication) and at test (sober, ascend-

ing limb, peak, descending limb). This design permitted us

to examine alcohol’s effects on memory through an ex-

panded, yet finely meshed, window.

A control group went through identical procedures

but under placebo conditions to control for factors such

as fatigue and interference from the massive amount of

materials.

Alcohol administration and BAC assessment

The alcohol group received 1 ml absolute alcohol/kg body

weight in a double-blind procedure, divided into three

drinks with 5 min given for the consumption of every

0.25 ml/kg of alcohol: 0.25 ml/kg at time 0–5 min

(immediately after SP2 and TP1; see Fig. 1a,b), 0.50 ml/kg

at time 45–55 min, and 0.25 ml/kg at time 90–95 min. Ten

minutes after each of the three drinks, subjects gargled with

water, and breath samples were taken using a Mark 4 Gas

Chromatographic Intoximeter. Alcohol drinks consisted of

one part 95% ethanol and seven parts peppermint-lemon-

ade masking solution. Masking solution was Minute Maid

Crystals at 150% recommended solution with 1.0 ml

peppermint extract per liter solution. The placebo group

received the same volume of masking solution with

0.25 ml alcohol floated on top of each drink to give

olfactory cues of alcohol. All participants were told they

might receive varying doses of alcohol during one of the

sessions or both. Although it is hard to simulate a high

alcohol dose, or to mask it, the possibility of receiving

alcohol, the alcohol smell of the placebo drink, and the

repetitive BAC readings contributed to increase expec-

tancy. The efficacy of the placebo condition was assessed

by asking participants to rate how “high” they felt (rang-

ing between 0=not at all high and 5=extremely high) and

how many drinks it would take to make them feel the

way they did (ranging between 0 and 12 or more drinks).

This was done repeatedly throughout both sessions, in

association with the mood questionnaire described below.

At each study period, BAC was assessed twice, the first

following the first memory task and the other preceding the

last. At test, BAC was assessed between tasks during TPs 2

low, 2 high, 4 high, and 4 low. TP3 covered all tasks and

was longer than TPs 2 and 4. BAC was therefore assessed

before, in the middle of, and after testing during this test

period.

Fig. 1 a Blood alcohol concentration during study. The figure

shows average BAC (error bars represent the standard deviations)

after the first and before the last memory task at each study period

(SP) of day 1. The three glasses represent alcohol intake, which

occurred at time 0, 45, and 90 min. b Blood alcohol concentration

during test. The figure shows average BAC (error bars represent the

standard deviations) assessed in the middle of each test period (TP)

of day 8. For TP3, BAC was also assessed before and after testing

(see text). The word fragment completion and free-recall tasks were

administered on the low rising and falling BAC (∼0.03 g/100 ml;

“TP2 low” and “TP4 low”), and the associative learning and picture

recognition tasks were administered on the high rising and falling

BAC (∼0.06 g/100 ml; “TP2 high” and “TP4 high”). The three

glasses represent alcohol intake, which occurred at time 0, 45, and

90 min

307

Memory tasks

Four different memory tasks were included to assay a wide

range of mnemonic processes. Tasks were selected to be

memorable a week later in the context of a vast amount of

other potentially interfering materials and sufficiently in-

teresting to engage the attention of intoxicated subjects

during two 9-h experimental sessions.

The materials of each memory task were organized into

32 subsets. Four subsets of materials were studied during

each of the eight study periods on day 1, and one subset

from each study period was tested during each of the four

test periods a week later. The 32 subsets were counter-

balanced across study and test periods. Detailed informa-

tion about the number of stimuli presented at each study

period and test period is shown in Table 1, and brief

descriptions of test procedures and administration time are

provided in Table 2.

Associative learning (recognition and cued recall) This

task is a variant of the traditional paired-associate task

typically used to assess associative learning. Each test

item consisted of a humorous unfamiliar “definition” of

a familiar noun, followed by the noun in question (e.g., a

wide sphere cracks its skin—EARTHQUAKE). The defi-

nition part of the test item serves as the “cue” at test, and

the noun as the “target” (i.e., material to remember) to first

be recognized and then recalled (Tulving and Watkins

1977). This task was included inasmuch as the unique

phrases have been shown to be memorable for a long time

after study (Hayman et al. 1993) and are frequently used in

cognitive psychology and neuropsychology. Half of the

test items were randomly categorized as cue–target pairs,

presented during both study and test, and half served as

distractor pairs (i.e., lures), presented only during test.

At study, subjects were asked to rate on a four-point

scale how much sense each definition made for the word

in question. At test, each previously encountered cue–

target pair, together with new similar distractors, was

tested twice, first for recognition from day 1 of the target

word without the cue (EARTHQUAKE—yes or no?), and

then for cued recall of the same target word (a vast sphere

cracks its skin—which word?). During cued recall, if

subjects were unable to explicitly recall the target, they

were encouraged to guess.

Recognition was scored in terms of “hits” (“yes”

responses to studied target words) and “false alarms”

(“yes” responses to nonstudied distractor words). The

difference between the proportions of hits and false alarms

reflects what subjects remember from the study list. A

subject’s cued recall score (proportion of correct target

words produced to cues), in this design, reflects the

combined effects of (1) associative learning retained over

7 days and (2) correct matching (“lucky guesses”) of the

target words, remembered from the preceding recognition

test, to their respective cues. Because such lucky guessing

is assumed to be comparable for the originally studied and

nonstudied items, the difference in the cued recall scores

between targets and distractors provides an estimate of

cued (associative) recall.

Picture recognition At study, subjects viewed complex

color photographs, selected from National Geographic and

from earlier studies on alcohol (cf. Parker et al. 1976,

1980; Tulving 1981). At test, subjects performed a forced-

choice recognition task where they chose which of two

slides they had seen on day 1. To make retrieval more

difficult, each target photograph was paired with a highly

similar distractor. The picture pairs were pilot tested to

ensure that the 32 subsets were equally difficult.

Word fragment completion (perceptual priming and rec-

ognition) This task consisted of completing single-solution

word fragments (e.g., A_ _ A _ _ IN). The materials

were words having three or four letters missing in their

most fragmented form. At both study and test, the frag-

ments were presented through ascending method of limits

(cf. Snodgrass and Feenan 1990), with missing letters

being successively added until the subject came up with

the word (e.g., ASSASSIN). Completion difficulty was

matched in the 32 subsets of materials based on pilot

testing. At test, subjects were tested for their ability to

complete the same fragments they completed during study

(targets) along with new fragments (distractors), and after

each completion, they were tested on their recognition of

the words generated from the fragments. Priming was

defined as the proportion of completed target fragments

on day 8 minus the proportion of completed distractor

fragments. Recognition was assessed in terms of hits and

false alarms.

Free recall Because the present design required materials

from certain study periods to be retrieved at certain test

periods, it precluded the use of a traditional free recall task

and called for a way to direct subjects what materials to

retrieve. Accordingly, a modified free recall task was

developed, with 24 words drawn from each of four

semantic categories (the 3rd to the 26th most frequent

Table 1 Number of stimuli in the memory tasks as a function of

study period, test period, and in total

Memory

task

Stimuli per

study period

Stimuli per

test period

Total Stimuli

TD T D

Associative learning 16 32 32 128 128

Picture recognition 24 48 48 192 192

Word fragment

completion

16 32 16 128 64

Free recall 12 24 NA 96 NA

All stimuli in a specific task were distributed across eight study

periods and four test periods. For example, for associative learning,

16 stimuli were presented during each of the eight study periods

(128 stimuli total). During each of the four test periods, 32 pre-

viously studied stimuli were presented as targets and 32 novel

stimuli were presented as distractors

T targets, D distractors, NA not applicable

308

category exemplars; Battig and Montague 1969). During

each study period, three words from each of the four

categories were presented (12 in total; see Table 1).

Subjects were asked to rank these words according to a

certain criterion (e.g., professions in terms of salary), and

thereafter write down the words. On day 8, during each of

the four test periods, one of the category names (counter-

balanced across test periods) was given to indicate what

words to retrieve. At each test period, participants could

thus recall three words from each of the eight study

periods, 24 words in total. The proportion of recalled

words was the measure of performance.

Associative learning and picture recognition were tested

at the high BAC, whereas perceptual priming and free

recall were tested at the low BAC. The rationale for this

allocation was that (1) the free recall task was expected to

be the most vulnerable to alcohol because it provides the

fewest retrieval cues, and it could therefore be tested at a

lower BAC than tasks less sensitive to alcohol; and (2)

tasks needed to be grouped so that each test period did not

exceed 30 min.

Profile of mood states

Mood was assessed throughout the study and test periods

(except at SP2) by means of the profile of mood states

(POMS) (McNair et al. 1971). This mood scale consists of

65 items where participants rate their state of mind (e.g.,

friendly, nervous) on a scale between 0 (not at all) and 4

(extremely). Six mood scores are obtained: tension, depres-

sion, anger, vigor, fatigue, and confusion. Most subjects

took about 3 min to complete the scale. Although this

activity constituted a diversion for the subjects, it was brief

and easy enough not to tax their resources.

Procedure

Subjects agreed to refrain from alcohol and other psycho-

active substances for 48 h and to fast from midnight

onward before each experimental session. They arrived at

the laboratory at 8 a.m. and were interviewed about com-

pliance with instructions. A baseline breath sample was

taken to ensure sobriety. Alcohol was given on an empty

stomach to reduce the variability in BAC among subjects.

Caloric intake was provided in the large volume of

concentrated lemonade masking solution. Subjects per-

formed the tasks in groups of two or four with an alcohol

subject and his yoked placebo subject going through the

sessions together. They brought their own lunch which they

were allowed to eat after SP5 of day 1 and TP3 of day 8.

Because alcohol is metabolized at a rate constant (Grilly

1998), differences in caloric content of the lunch should not

cause variability in BAC during the study and test periods

after peak. Breath samples were taken at the end of both

Table 2 Procedure and distribution time of memory measures at study and test

Memory measure Procedure at study Distribution time at

study (per item and

in total)

Procedure at test Distribution time at test

(per item and in total)

Associative learning

(recognition/cued

recall)

Read phrase–word pairs and

rate how much sense they make

8 sec/pair;

3 min total

Recognition: judge

whether a presented

word is old or new

Recognition: 5 sec/word;

6 min total

Cued recall: come up

with a word given the

phrase

Cued recall: 8 sec/phrase;

9 min total

Picture recognition Watch photos 5 sec/photo;

2.5 min total

Chose from two

similar photos which

one was seen before

10 sec/pair; 9 min total

Word-fragment

completion

(priming/

recognition)

Complete word fragments 8 sec/fragment slide;

11 min total

Priming: complete word

fragments

8 sec/fragment slide;

27 min total

Recognition: judge

whether completed

word is old or new

Free recall Rank words according to certain

semantic criteria followed by

immediate recall

10 sec/triad;

1 min total

Recall words given

the category cue,

one category per test

period

3 min per category;

3 min total (i.e., one

category tested per test

period)

Total distribution times include instructions

309

sessions to verify sobriety. The sessions lasted approxi-

mately 9 h, and subjects left the laboratory when BAC had

reached zero.

Analyses

The eight study periods and four test periods gave rise to 32

measurement points. Mixed-design ANOVA was used for

all analyses, with group as between-subjects factor (alcohol

vs placebo) and study and test periods as within-subjects

factors. Depending on the issue under examination, dif-

ferent points of the 8×4 study and test period matrix were

included. A detailed description of the specific analyses

is provided for the recognition part of the associative

learning task, and the same analyses were performed on

all the other tasks. No adjustment was made for initial

performance, because the groups had equivalent sober

peformance (see SP1/TP1 in Tables 3, 4, 5), tested by

t tests.

Results

Alcohol expectancy

The alcohol group felt significantly “higher” than the

placebo group (assessed through t tests, ps≤0.003) at study

periods 4 through 7 and test periods 2 low through 4 low.

At study, the alcohol group’s high scores were 2.2, 2.3, 1.8,

and 0.7, respectively, whereas the placebo group’s high

scores were 0.9, 0.5, 0.2, and 0.1, respectively. At test, the

alcohol group’s high scores were 1.4, 2.5, 2.1, 1.0, and 0.7

whereas the placebo group had scores of 0.3, 0.5, 0.2, 0.2,

and 0.1. Although the differences in high were significant

between the two groups, the placebo group’s average high

scores were all above zero. The alcohol group estimated

they had had more to drink than the placebo group at study

periods 3 through 6, and test periods 2 high through 4 low

(ps≤0.006). Nevertheless, through study periods 3 to 6, 47,

50, 28, and 6%, respectively, of the subjects in the placebo

group believed they had had some alcohol (ranging

between one and four drinks). At test, these values were

16, 16, 6, and 6%, respectively.

Associative learning

Recognition (EARTHQUAKE?)

Overall effects of alcohol Alcohol’s overall effects were

assessed using mixed design 2 (Group) ×5 (SP 3–7) ×3

(TP 2–4) ANOVAs (see the values in bold in Table 3 and

corresponding tables for the other memory measures).

These study and test periods were included because the

alcohol group and the placebo group differed in BAC

during these periods. The number of recognition hits

showed a tendency of an overall effect of alcohol

(1,62)

=3.51, p=0.07] where the alcohol group generally

Table 3 Recognition of words in the associative learning task, Mean (SD) proportion of hits and false alarms in the placebo and alcohol groups as a function of study and test periods and

the BACs (g/100 ml) in the alcohol group corresponding to that study or test period

Test period

and BAC

Study period and BAC

SP1 SP2 SP3 SP4 SP5 SP6 SP7 SP8 Mean hits

(test)

Mean false

alarms0.00 0.00 0.03 asc. 0.06 asc. 0.08 peak 0.06 desc. 0.03 desc. 0.00

TP1 0.00 P 0.66 (0.23) 0.66 (0.30) 0.65 (0.25) 0.68 (0.25) 0.56 (0.27) 0.66 (0.25) 0.63 (0.27) 0.75 (0.24) 0.66 (0.16) 0.39 (0.16)

A 0.73 (0.26) 0.69 (0.26) 0.72 (0.24) 0.66 (0.27) 0.66 (0.22) 0.66 (0.24) 0.75 (0.22) 0.73 (0.27) 0.70 (0.14) 0.43 (0.16)

TP2 0.06 asc. P 0.74 (0.24) 0.80 (0.21)* 0.77 (0.28)* 0.74 (0.23)§ 0.75 (0.22)* 0.69 (0.27) 0.66 (0.24) 0.65 (0.26) 0.72 (0.15)† 0.41 (0.19)

A 0.69 (0.26) 0.64 (0.32) 0.59 (0.27) 0.50 (0.30) 0.61 (0.28) 0.60 (0.31) 0.57 (0.28) 0.70 (0.26) 0.61 (0.18) 0.39 (0.12)

TP3 0.08 peak P 0.72 (0.27) 0.67 (0.32) 0.72 (0.30) 0.72 (0.26) 0.66 (0.30) 0.70 (0.27) 0.68 (0.27) 0.70 (0.26) 0.70 (0.19) 0.42 (0.22)

A 0.63 (0.31) 0.70 (0.25) 0.60 (0.28) 0.66 (0.24) 0.59 (0.23) 0.63 (0.27) 0.64 (0.28) 0.67 (0.26) 0.64 (0.16) 0.44 (0.15)

TP4 0.06 des. P 0.68 (0.25) 0.70 (0.28) 0.65 (0.31) 0.63 (0.32) 0.66 (0.25) 0.63 (0.32) 0.70 (0.21) 0.63 (0.32) 0.66 (0.18) 0.42 (0.20)

A 0.81 (0.21)* 0.67 (0.29) 0.70 (0.25) 0.66 (0.27) 0.59 (0.29) 0.64 (0.30) 0.64 (0.31) 0.63 (0.29) 0.67 (0.19) 0.46 (0.17)

Mean hits (study) P 0.70 (0.18) 0.71 (0.20) 0.70 (0.19) 0.69 (0.19) 0.66 (0.18) 0.67 (0.18) 0.67 (0.17) 0.68 (0.19) 0.68 (0.15) 0.41 (0.17)

A 0.72 (0.20) 0.67 (0.20) 0.65 (0.17) 0.62 (0.20) 0.61 (0.17) 0.63 (0.19) 0.65 (0.20) 0.68 (0.19) 0.66 (0.25) 0.43 (0.12)

False alarms are only reported as a function of test period inasmuch as distractors were only presented at test. The values in bold indicate data used in analyses of main effects of alcohol

P placebo, A alcohol, asc. ascending limb, desc. descending limb

*p<0.05; †p<0.01; §p=0.001. The significance symbols mark the group that had the significantly higher performance

310

Table 4 Cued recall in the associative learning task, Mean (SD) proportion of correctly solved targets and distractors in the placebo and alcohol groups as a function of study and test

periods and the BACs (g/100 ml) in the alcohol group corresponding to that study or test period

Test period

and BAC

Study period and BAC

SP1 SP2 SP3 SP4 SP5 SP6 SP7 SP8 Mean solved

targets (test)

Mean solved

distractors

Correct. mean

solved targets0.00 0.00 0.03 asc. 0.06 asc. 0.08 peak 0.06 desc. 0.03 desc. 0.00

TP1 0.00 P 0.38 (0.23) 0.41 (0.30) 0.46 (0.24) 0.36 (0.28) 0.34 (0.22) 0.34 (0.30) 0.41 (0.29) 0.36 (0.30) 0.38 (0.16) 0.12 (0.08) 0.26 (0.14)*

A 0.30 (0.26) 0.37 (0.24) 0.35 (0.31) 0.32 (0.26) 0.23 (0.23) 0.27 (0.29) 0.31 (0.26) 0.34 (0.27) 0.31 (0.17) 0.13 (0.08) 0.19 (0.14)

TP2 0.06 asc. P 0.39 (0.25)* 0.44 (0.32)* 0.41 (0.28)† 0.33 (0.27) 0.34 (0.30)† 0.34 (0.26)† 0.31 (0.28) 0.38 (0.26) 0.37 (0.16)‡ 0.13 (0.08)‡ 0.24 (0.15)*

A 0.26 (0.22) 0.27 (0.26) 0.24 (0.22) 0.23 (0.23) 0.16 (0.20) 0.17 (0.18) 0.20 (0.23) 0.28 (0.23) 0.23 (0.12) 0.06 (0.05) 0.17 (0.10)

TP3 0.08 peak P 0.33 (0.27) 0.37 (0.30) 0.35 (0.25) 0.33 (0.25)* 0.27 (0.24)* 0.41 (0.23)§ 0.36 (0.24)* 0.41 (0.30)§ 0.35 (0.15)‡ 0.14 (0.09)§ 0.22 (0.12)*

A 0.23 (0.17) 0.23 (0.25) 0.34 (0.27) 0.20 (0.21) 0.16 (0.18) 0.18 (0.22) 0.23 (0.26) 0.18 (0.22) 0.22 (0.12) 0.07 (0.06) 0.15 (0.10)

TP4 0.06 des. P 0.37 (0.25) 0.39 (0.25) 0.41 (0.30)* 0.39 (0.25)* 0.38 (0.25)† 0.35 (0.26)† 0.33 (0.27) 0.44 (0.25)† 0.38 (0.14)‡ 0.14 (0.09)* 0.24 (0.11)†

A 0.26 (0.20) 0.31 (0.21) 0.27 (0.22) 0.24 (0.24) 0.22 (0.21) 0.17 (0.21) 0.24 (0.27) 0.25 (0.28) 0.25 (0.13) 0.09 (0.07) 0.16 (0.10)

Mean solved

targets (study)

P 0.37 (0.18)* 0.40 (0.21)* 0.41 (0.16)† 0.35 (0.17)* 0.33 (0.16)‡ 0.36 (0.15)‡ 0.35 (0.16)* 0.40 (0.18)† 0.37 (0.14)‡ 0.13 (0.06)§ 0.24 (0.10)§

A 0.26 (0.14) 0.30 (0.14) 0.30 (0.16) 0.25 (0.16) 0.19 (0.14) 0.20 (0.15) 0.25 (0.16) 0.26 (0.16) 0.25 (0.11) 0.08 (0.05) 0.17 (0.08)

The values in bold indicate data used in analyses of main effects of alcohol

P placebo, A alcohol, asc. ascending limb, desc. descending limb, Correct. corrected

*p<0.05; †p<0.01; §p<0.001; ‡ p<0.0001. The significance symbols mark the group that had the significantly higher performance

Table 5 Recognition of completed word fragments, Mean (SD) proportion of hits and false alarms in the placebo and alcohol groups as a function of study and test periods and the BACs

(g/100 ml) in the alcohol group corresponding to that study or test period

Test period

and BAC

Study period and BAC

SP1 SP2 SP3 SP4 SP5 SP6 SP7 SP8 Mean hits

(test)

Mean false

alarms0.00 0.00 0.03 asc. 0.06 asc. 0.08 peak 0.06 desc. 0.03 desc. 0.00

TP1 0.00 P 0.91 (0.15) 0.88 (0.17) 0.91 (0.18) 0.90 (0.18) 0.89 (0.18)* 0.89 (0.15)* 0.91 (0.14) 0.91 (0.16) 0.90 (0.10) 0.17 (0.09)

A 0.91 (0.15) 0.90 (0.17) 0.87 (0.23) 0.83 (0.20) 0.78 (0.21) 0.80 (0.21) 0.86 (0.18) 0.88 (0.16) 0.85 (0.13) 0.15 (0.08)

TP2 0.03 asc. P 0.92 (0.12) 0.93 (0.13) 0.91 (0.14) 0.91 (0.14)† 0.90 (0.15)* 0.92 (0.13)* 0.91 (0.16) 0.92 (0.13) 0.92 (0.07)* 0.21 (0.12)

A 0.92 (0.13) 0.88 (0.18) 0.88 (0.22) 0.77 (0.20) 0.76 (0.26) 0.83 (0.21) 0.84 (0.19) 0.89 (0.15) 0.85 (0.13) 0.17 (0.09)

TP3 0.08 peak P 0.94 (0.13)* 0.87 (0.22) 0.88 (0.18) 0.90 (0.18) 0.88 (0.18)* 0.84 (0.18) 0.89 (0.15) 0.92 (0.13)* 0.89 (0.09)* 0.23 (0.11)

A 0.83 (0.22) 0.86 (0.23) 0.83 (0.21) 0.83 (0.22) 0.76 (0.25) 0.77 (0.29) 0.84 (0.20) 0.80 (0.24) 0.81 (0.17) 0.18 (0.08)

TP4 0.03 des. P 0.91 (0.22) 0.88 (0.17) 0.87 (0.17) 0.95 (0.12)† 0.88 (0.14) 0.86 (0.20) 0.87 (0.19) 0.88 (0.18) 0.89 (0.09) 0.24 (0.12)

A 0.91 (0.19) 0.88 (0.22) 0.83 (0.25) 0.83 (0.20) 0.81 (0.23) 0.82 (0.24) 0.86 (0.17) 0.91 (0.16) 0.86 (0.13) 0.21 (0.09)

Mean hits (study) P 0.92 (0.08) 0.89 (0.14) 0.89 (0.10) 0.91 (0.11)§ 0.89 (0.11)§ 0.88 (0.11) 0.89 (0.10) 0.91 (0.10) 0.90 (0.08)* 0.21 (0.10)

A 0.89 (0.13) 0.88 (0.16) 0.85 (0.17) 0.81 (0.13) 0.78 (0.17) 0.80 (0.18) 0.85 (0.13) 0.87 (0.12) 0.84 (0.13) 0.18 (0.06)

The values in bold indicate data used in analyses of main effects of alcohol

P placebo, A alcohol, asc. ascending limb, desc. descending limb

*p<0.05; †p<0.01; §p<0.005. The significance symbols mark the group that had the significantly higher performance

311

had fewer hits than the placebo group (11% fewer hits on

average; see Table 3). There was a significant interaction

between group and test period [F

(2,124)

=4.95, p=0.01] due

to the alcohol group performing worse than the placebo

group on the ascending limb (see Table 3, TP2), but not at

the other test periods.

As distractors were presented only at test, false alarms

were analyzed in terms of group and test by mixed design

2 (Group) ×3 (TP 2–4) ANOVAs. The proportion of false

alarms did not differ between groups (F<1), so there was

no need for correction of the hits.

Simple effects of alcohol on encoding and retrieval To

assess alcohol’s effect on encoding without potential con-

tamination by its effect on retrieval, intoxicated encoding

(SP 3–7) and sober retrieval (TP1) were analyzed using a

mixed design 2 (Group) ×5 (SP 3–7) ANOVA. In a similar

fashion, only sober encoding (SP1) was included to study

alcohol’s pure effect on retrieval, at TPs 2 to 4 [2 (Group)

×3 (TP 2–4) repeated measures ANOVA]. SP2 was not

included because of possible memory enhancement of

material encoded just prior to alcohol intake (cf. Parker

et al. 1980, 1981). There was no simple effect of alcohol

on encoding [F

(1,62)

=1.06, ns] or retrieval (F<1). How-

ever, group and TP interacted [F

(2,124)

=5.64, p=0.005]

due to the alcohol group performing significantly better

than the placebo group on the descending limb (see

Table 3, SP1, TP4). Because the alcohol group had higher

performance than the placebo group only at this mea-

surement point out of 32 possible points, and because this

finding is in opposition with previous research, it may

be a spurious finding.

Effects of alcohol on the ascending and descending limbs

The marginal means of SPs 3, 4, 6, and 7 and TPs 2 and 4

were included in 2 (Group) ×2 (Limb) mixed-design

ANOVAs to compare the effect of alcohol for low and

high BACs on the ascending vs. descending limbs at study

(SP3 vs. SP7 and SP4 vs. SP6, respectively) and the

ascending vs. descending BAC at retrieval (TP2 vs. TP4).

An interaction between group (alcohol vs. placebo) and

limb (ascending vs. descending) would suggest a differ-

ential effect of alcohol on the ascending and descending

limbs of the BAC curve. There was no interaction between

limb and group at study at either the high or low BAC

(Fs<1), but the two factors interacted at retrieval

(1,62)

=10.43, p<0.005]. This interaction was due to the

alcohol group having fewer hits than the placebo group

on the ascending limb at retrieval. There was no in-

teraction between limb and group in terms of false alarms

(1,62)

=2.85, p =0.10].

Cued recall (a vast sphere cracks it skin—_______?)

Overall effects of alcohol Identical analyses as those

described above revealed a significant overall effect of

alcohol on cued recall of target words [F

(1,62)

=20.42,

p<0.0001]. As can be seen in Table 4, the alcohol group

performed worse than the placebo group throughout all

study periods and at test at rising, peak, and falling BAC

(39% fewer correctly recalled words on average).

Alcohol also had an overall effect on solving distractor

cues (lucky guesses) [F

(1,62)

=19.71, p<0.0001] that were

only seen on day 8. Impairment occurred at all test periods

where alcohol was present (i.e., TPs 2–4 in Table 4). To

get an estimate of pure memory performance that does not

comprise the ability to solve the phrase cues through

semantic elaboration and/or guessing, the proportions of

solved distractor cues were subtracted from the propor-

tions of solved target cues (e.g., the “mean solved dis-

tractors” value of TP1 was subtracted from the proportions

of solved targets of TP1/SPs 3–7, respectively). Having

corrected for the general ability to solve distractors, the

effect of alcohol on cued recall remained significant

(1,62)

=11.07, p=0.001], although smaller, partial η

=0.15

as compared to partial η

=0.25 for uncorrected solved

targets.

Simple effects of alcohol on encoding (sober retrieval) and

retrieval (sober encoding) Alcohol impaired encoding of

materials that were recalled while sober [F

(1,62)

=4.01,

p=0.05]. Alcohol also impaired retrieval when study had

been sober [F

(1,62)

=7.43, p<0.01], but this effect was no

longer significant after correction for distractor solving

(1,61)

=1.94, ns]. The retrieval effect of alcohol was thus

mainly due to alcohol’s detrimental effect on solving cues

in general.

Effects of alcohol on the ascending and descending limbs

There were no interactions between group and limb either

at encoding or retrieval in corrected target solving (Fs<1).

Picture recognition

There was no overall effect of alcohol on picture recog-

nition, nor were there any effects on encoding, retrieval, or

any interactions (Fs<1).

Word fragment completion

Perceptual priming The overall completion rates of target

and distractor fragments on day 8 were 0.21 and 0.09 in

the alcohol group and 0.22 and 0.08 in the placebo group,

giving rise to respective priming scores of 0.12 and 0.14.

Alcohol had no overall effect on the amount of priming,

and no effects were observed in any of the focal analyses

(Fs≤1).

Recognition of completed word fragments

Overall effects of alcohol Alcohol reduced the number of

hits [F

(1,62)

=7.16, p<0.05], with the alcohol group always

having fewer hits than the placebo group (8% fewer hits

on average; see Table 5). The rate of false alarms was not

affected by alcohol [F

(1,62)

=2.87, p<0.05], so there was no

need to adjust hits for false alarms.

312

Simple effects of alcohol on encoding (sober retrieval) and

retrieval (sober encoding) Recognition hits were signifi-

cantly reduced when subjects encoded under alcohol and

retrieved when sober [F

(1,62)

=5.22, p<0.05]. Although

alcohol had no main effect on retrieval, group interacted

with test period [F

(2,124)

=3.77, p<0.05] due to the alcohol

group having fewer hits at ascending and peak BAC at test

(TPs 2 and 3; see Table 5).

Effects of alcohol on the ascending and descending limbs

There was no interaction between group and limb at

study (F<1), but at retrieval, group and limb interacted

(1,62)

=4.31, p<0.05] due to the alcohol group having

fewer hits than the placebo group on the ascending but

not the descending limb. There was no interaction in terms

of false alarms between group and limb (F<1).

Free recall

Overall effects of alcohol Alcohol had an overall effect

on the number of words recalled [F

(1,62)

=5.97, p<0.05,],

globally reducing performance (the average recall was

0.25±0.07 in the alcohol group and 0.30±0.10 in the

placebo group).

Simple effects of alcohol on encoding (sober retrieval)

and retrieval (sober encoding) There was a significant

detrimental effect of alcohol on encoding [F

(1,62)

=6.33,

p<0.05], but not on retrieval (F<1). Under conditions of

intoxicated encoding and sober retrieval, the alcohol group’s

average recall was 0.20±0.11, compared to the placebo

group’s average recall of 0.27±0.10 at the same study and

test periods. Under conditions of sober encoding and

intoxicated retrieval, the groups’ recall was 0.34±0.16 and

0.34±0.20, respectively.

Effects of alcohol on the ascending and descending limbs

There was no interaction between group and limb either at

encoding or retrieval.

Mood There was no difference between groups in sober

mood scores (tested by t tests), so no adjustment was made

for baseline mood. At encoding, 2 (Group) ×2 (Limb)

mixed-design ANOVAs run separately for low and high

BAC revealed differential effects of alcohol on the as-

cending and descending limbs in the depression, vigor, and

fatigue scores as reflected in group × limb interactions (see

Fig. 2, left column). At identical BAC (∼0.03 g/100 ml),

the alcohol group’s depression scores were higher on the

descending limb than on the ascending limb, in contrast to

the placebo group which showed the reverse pattern. The

alcohol group’s vigor scores were lower on the descending

limb than on the ascending limb, whereas there was no

difference between limbs in the placebo group. The op-

posite was true for the fatigue scores, which were higher in

the alcohol group on the descending limb (low and high

BAC) than on the ascending limb, but were similar in

the placebo group. Generally, the alcohol group had some-

what higher scores on tension, fatigue, and confusion. At

retrieval, no interactions between group and limb were

observed, although the alcohol group generally had some-

what higher tension, anger, and confusion scores than

the placebo group, and particularly so on the high as-

cending limb (see Fig. 2, right column).

Discussion

The present study was undertaken to examine alcohol’s

effects on memory encoding and retrieval during the

ascending and descending limbs of the BAC curve using a

new design. The experiment included multiple memory

tasks distributed at encoding and retrieval throughout the

whole BAC curve when subjects had consumed alcohol

gradually, as may occur in social drinking situations.

Further, recall took place one week after study, which is a

longer retention interval than that used in most studies,

providing an estimate of alcohol’s effects on long-term

memory. Earlier studies have contrasted effects of alco-

hol on the ascending and descending limbs on different

cognitive functions and mood, but this is the first study

to examine the effects of cumulative drinking on memory

throughout the whole intoxication period, during both

encoding and retrieval.

In line with previous research and our expectations,

alcohol impaired episodic encoding. This was true for cued

recall, recognition of completed word fragments, and free

recall. Unexpectedly, alcohol also impaired retrieval in

two word recognition tasks (in associative learning and

word fragment completion), particularly on the ascending

limb of the BAC curve. Previous findings of alcohol-

resistant forms of memory, specifically perceptual priming,

were replicated in this study, but a lack of alcohol effect

was also found in episodic memory for complex pictorial

scenes. Contrary to our predictions, the memory measure

most impaired by alcohol was cued recall rather than free

recall, with the largest effects occurring when both en-

coding and retrieval took place during intoxication.

Alcohol impaired encoding in three verbal episodic

memory tasks (cued recall, recognition of completed word

fragments, and free recall). Impaired encoding of verbal

materials under alcohol has been observed before with

slightly higher alcohol doses (Goodwin et al. 1969;

Petersen 1977). Research in rodents has shown that alcohol

distorts the function of the hippocampus and the septo-

hippocampal pathway (Givens 2000; White 2000) and

blocks hippocampal long-term potentiation (Sinclair and

Lo 1986; Givens and McMahon 1995). An alteration of

hippocampal function may in part be the cause of the

reduced encoding and, to a smaller extent, retrieval, given

the hippocampus’ differential involvement in these two

processes (Lepage et al. 1998). Episodic encoding and

retrieval also differentially involve the prefrontal cortex

(Tulving et al. 1994), which is yet another structure to

explore in relation to alcohol-induced memory impairment.

However, inferences on the locus of alcohol’s impact on

the brain made on the basis of behavioral data remain

313

speculative and should be examined by means of func-

tional neuroimaging.

Although previous research has found that encoding

is more impaired than retrieval (Goodwin et al. 1969;

Petersen 1977; Birnbaum et al. 1978; but see Fillmore

et al. 1999), it is not clear in what phase of the BAC

curve retrieval took place. Using a slowly rising BAC,

our study showed alcohol effects also at retrieval, during

the ascending limb. Limb of the BAC curve is thus an

important dimension in delineating alcohol’s amnesic ef-

fects. One explanation of impaired retrieval on the as-

cending but not descending limb is the development of

acute tolerance with alcohol (cf. the Mellanby effect;

Mellanby 1919; Moskowitz et al. 1979) on the descend-

ing limb. No such tolerance was observed at encoding,

where the alcohol effect was more general and not re-

Fig. 2 Average POMS sub-

scores during the different study

periods (left) and test periods

(right) in the alcohol group (

▪

)

and the placebo group (□ ). The

error bars represent the standard

deviations and are only shown

in one direction for clarity.

lo=low, hi=high. Score differ-

ences between groups were

assessed by t tests: *p<0.05;

†p<0.005; ‡p<0.0001;

Group ×

Limb interaction F

(1,60)

=4.36,

p<0.05;

Group × Limb inter-

action F

(1,61)

=12.84, p<0.005;

Group × Limb interaction

(1,60)

=8.62, p<0.01;

Group ×

Limb interaction F

(1,58)

=21.43,

p<0.0001;

Group × Limb in-

teraction F

(1,60)

=6.78, p<0.05

314

stricted to a particular limb. This could imply that the

effect at encoding was of such strength that acute tol-

erance could not be developed. In a study on acute tol-

erance and motor performance, tolerance was more

frequently developed in the group that received the

lower out of two-alcohol doses (0.75 g/kg as compared to

1.0 g/kg; Bennett et al. 1993). Hence, improvement in

performance on the descending limb as compared to the

ascending limb likely occurs when sufficient mental re-

sources are available due either to a relatively low BAC

(e.g., 0.75 g/kg) or to a task or process that is relatively

little impaired by alcohol (e.g., retrieval).

The neural correlates of differential behavior on the

ascending and descending limbs have only been explored

in a few studies. Using positron emission topography

(PET), activation differences between the two limbs were

found in right prefrontal cortex, anterior cingulate, and left

superior temporal cortex (Schreckenberger et al. 2004). A

reduced and delayed P300 in an auditory oddball paradigm

during the ascending BAC was revealed by means of

event-related potentials (ERPs; Lukas et al. 1990), possibly

reflecting a disturbance of such functions as attention,

memory updating, or information transfer to controlled

processing and consciousness (Picton 1992). However, no

clear comparison was made between the ascending and

descending limbs in the aforementioned study (Lukas et al.

1990). Further exploration of cerebral activity and behavior

during the two limbs should elucidate the differential

effects on memory during the ascending and descending

limbs of the BAC curve.

Alcohol’s differential effect during the two limbs

extended to mood. During the study session, the alcohol

group scored higher on fatigue, depression, and confusion

and lower on vigor on the descending limb than on the

ascending limb, confirming earlier findings of negative

mood during the descending limb (Babor et al. 1983;

Earleywine and Martin 1993; Sutker et al. 1983). However,

at test one week later, no mood difference between the

ascending and descending limbs was found, although the

alcohol group felt more tense, depressed, angry, and con-

fused than the placebo group on the ascending limb at

the high BAC. It is not clear why the mood response

to alcohol would differ between the two sessions. The

POMS is usually a reliable tool at longer time intervals

when there is no intervention (Salinsky et al. 2001), but it

is possible that having a second experimental session so

close in time to the first made participants react differ-

ently to the alcohol during the second session.

Contrary to our expectations, in spite of specific retrieval

cues being provided, cued recall was more impaired by

alcohol than all other tasks, including free recall. This

finding may be due to both free recall being reduced in

both groups after the weeklong retention interval and to the

elaborative processing of learning new semantic associa-

tions required in the cued recall task being impaired by

alcohol. One of alcohol’s effects is to reduce this kind of

processing (Hashtroudi et al. 1984; Maylor et al. 1987),

and deeper processing during intoxicated encoding does

not benefit subsequent recall to the same extent as it does

under sober conditions (Curran and Hildebrandt 1999;

Hartley et al. 1978). The differential long-term effects of

alcohol on the two tasks should be confirmed in future

research.

The results from this study reveal the importance of

the memory task itself in describing alcohol’s effects on

memory. For example, word-fragment completion (prim-

ing) was resistant to alcohol, replicating previous findings

(Fillmore et al. 1999; Hashtroudi et al. 1984; Nilsson et al.

1989; Ray et al. 2004; Tracy and Bates 1999), whereas

recognition of the completed fragments was significantly

impaired by alcohol overall and by alcohol at encoding.

Similarly, there were very different effects of alcohol on

the two memory measures in the associative learning

task. Recognition was sensitive to alcohol’s detrimental

effects on retrieval, particularly during ascending BAC.

Cued recall, on the other hand, was sensitive to alcohol in

general and to alcohol’s detrimental effects on encod-

ing. Although cross-task comparisons are limited to those

tasks tested at similar BACs, very different results were

thus obtained between and within tasks tested at the same

BACs, demonstrating the selectivity of alcohol’s effects

on memory.

The weeklong retention interval used here limits com-

parison to previous research where the typical design

has retrieval closely follow encoding on the same day.

Although relevant for understanding the long-term effects

of alcohol on memory and separating between alcohol

effects on encoding and retrieval, the 1-week delay may

have attenuated some of alcohol’s effects due to forgetting

in both groups. This possibility is supported by the results

of the Parker et al. (1976) study that showed alcohol-

induced retention decrement in immediate testing, but not 2

weeks later.

The exclusion of women in the present study also limits

the findings to men. Apart from concerns of giving alcohol

to women that may be undetectably pregnant, only men

were selected to reduce variability of the data, inasmuch as

it is known that cognitive skills can be affected differ-

entially by alcohol in the two genders (e.g., Mumenthaler

et al. 1999).

The mere expectation of alcohol frequently alters social

behavior (Hull and Bond 1986; Marlatt and Rohsenow

1980) and has varying effects on cognitive and motor

function. Such expectancy can improve performance by

means of compensation (Marczinski and Fillmore 2005),

but can also impair performance (Fillmore et al. 1998)or

have no noticeable effect (Connors and Maisto 1980

;

Rimm et al. 1982; Nagoshi et al. 1992). Memory per-

formance is usually unaffected by alcohol expectancy

(Miller et al. 1978; Nelson et al. 1986; Assefi and Garry

2003), although memory was reduced as compared to

controls following the ingestion of a placebo pill said to

impair memory performance (Kvavilashvili and Ellis

1999). The absence in previous research of alcohol’s ex-

pectancy effects on memory, and the successful placebo

deception in the present study, caused us to limit our

analyses to the direct effects of actual blood alcohol levels.

315

Although the current analyses were restricted to certain

points of the BAC curve according to the question under

investigation, using this type of paradigm also provides

data tailored well to show the combined effects of hav-

ing different BACs at encoding and retrieval and offer

information on such topics as retrograde facilitation, state-

dependent retrieval, and hangover effects for those

wishing to explore these areas.

To summarize, the results of this study provide support

for further investigation of alcohol’s effects on human

memory with reference to (1) distinguishing alcohol’s ef-

fects on encoding vs retrieval, (2) limb of the BAC curve

with particular attention to the ascending limb, (3) the

nature of the memory probe, and (4) the critical interactions

among these factors. These key findings may be useful in

guiding future research on the neurobiological substrates of

alcohol’s effects on human memory.

Acknowledgements Endel Tulving’s research is supported by the

Natural Sciences and Engineering Research Council of Canada and

by an endowment by Anne and Max Tanenbaum in support of

research in cognitive neuroscience. Data collection was supported by

the National Institute on Alcohol Abuse and Alcoholism.

References

Assefi SL, Garry M (2003) Absolut memory distortions: alcohol

placebos influence the misinformation effect. Psychol Sci 14:

77–80

Babor TF, Berglas S, Mendelson JH, Ellingboe J, Miller K (1983)

Alcohol, affect, and the disinhibition of verbal behavior.

Psychopharmacology 80:53–60

Battig WF, Montague WE (1969) Category norms for verbal items

in 56 categories. A replication and extension of the Connecticut

category norms. J Exp Psychol Mon 80:1–46

Bennett RH, Cherek DR, Spiga R (1993) Acute and chronic alcohol

tolerance in humans: effects of dose and consecutive days of

exposure. Alcohol Clin Exp Res 17:740–745

Birnbaum IM, Parker ES, Hartley JT, Noble EP (1978) Alcohol and

memory: retrieval processes. J Verbal Learn Verbal Behav

17:325–335

Connors G, Maisto S (1980) Effects of alcohol, instructions, and

consumption rate on motor performance. J Stud Alcohol 41:

509–517

Conrod PJ, Peterson JB, Pihl RO, Mankowski S (1997) Biphasic

effects of alcohol on heart rate are influenced by alcoholic

family history and rate of alcohol ingestion. Alcohol Clin Exp

Res 21:140–149

Curran HV, Hildebrandt M (1999) Dissociative effects of alcohol on

recollective experience. Conscious Cogn 8:497–509

Duka T, Weissenborn R, Dienes Z (2001) State-dependent effects of

alcohol on recollective experience, familiarity and awareness of

memories. Psychopharmacology 153:295–306

Earleywine M, Martin CS (1993) Anticipated stimulant and sedative

effects of alcohol vary with dosage and limb of the blood

alcohol curve. Alcohol Clin Exp Res 17:135–139

Eckardt MJ, File SE, Gessa GL, Grant KA, Guerri C, Hoffman PL,

Kalant H, Koob GF, Li TK, Tabakoff B (1998) Effects of

moderate alcohol consumption on the central nervous system.

Alcohol Clin Exp Res 22:998–1040

Erblich J, Earleywine M (2003) Behavioral undercontrol and

subjective stimulant and sedative effects of alcohol intoxica-

tion: independent predictors of drinking habits? Alcohol Clin

Exp Res 27:44–50

Fillmore MT, Carscadden JL, Vogel-Sprott M (1998) Alcohol,

cognitive impairment and expectancies. J Stud Alcohol 62:174–

179

Fillmore MT, Vogel-Sprott M, Gavrilescu D (1999) Alcohol effects

on intentional behavior: dissociating controlled and automatic

influences. Exp Clin Psychopharmacol 7:372–378

Fillmore MT, Dixon MJ, Schweizer TA (2000) Differential effects of

alcohol on responses to negatively and positively primed

stimuli. J Stud Alcohol 61:872–880

Giancola PR, Zeichner A (1997) The biphasic effects of alcohol on

human physical aggression. J Abnorm Psychol 106:598–607

Givens B (2000) Septohippocampal pathway as a site for the mem-

ory-impairing effects of ethanol. Hippocampus 10:111–121

Givens B, McMahon K (1995) Ethanol suppresses the induction of

long-term potentiation in vivo. Brain Res 688:27–33

Goodwin DW, Powell B, Bremer D, Hoine H, Stern J (1969)

Alcohol and recall: state-dependent effects in man. Science

163:1358–1360

Grilly DM (1998) Drugs and human behavior, 3rd edn. Allyn and

Bacon, Needham Heights, Massachusetts

Hartley JT, Birnbaum I, Parker ES (1978) Alcohol and storage

deficits: kind of processing? J Verbal Learn Verbal Behav 17:

635–647

Hartocollis P, Johnson DM (1956) Differential effects of alcohol on

verbal fluency. Q J Stud Alcohol 17:183–189

Hashtroudi S, Parker ES, DeLisi LE, Wyatt RJ, Mutter SA (1984)

Intact retention in acute alcohol amnesia. J Exp Psychol Learn

Mem Cogn 10:156–163

Hayman CAG, MacDonald CA, Tulving E (1993) The role of

repetition and associative interference in new semantic learning

in amnesia—a case experiment. J Cogn Neurosci 5:375–389

Hiltunen AJ (1997) Acute alcohol tolerance in cognitive and

psychomotor performance: influence of the alcohol dose and

prior alcohol experience. Alcohol 14:125–130

Holdstock L, de Wit H (1998) Individual differences in the biphasic

effects of ethanol. Alcohol Clin Exp Res 22:1903–1911

Hull JG, Bond CF Jr (1986) Social and behavioral consequences of

alcohol consumption and expectancy: a meta-analysis. Psychol

Bull 99:347–360

Hurst PM, Bagley SK (1972) Acute adaptation to the effects of

alcohol. Q J Stud Alcohol 33:358–378

Ilan AB, Gevins A (2001) Prolonged neurophysiological effects of

cumulative wine drinking. Alcohol 25:137–152

Jones BM (1973) Memory impairment on the ascending and

descending limbs of the blood alcohol curve. J Abnorm Psychol

82:24–32

Jones BM, Vega A (1972) Cognitive performance measured on the

ascending and descending limbs of the blood alcohol curve.

Psychopharmacology 23:99–114

King A, Houle T, de Wit H, Holdstock L, Schuster A (2002)

Biphasic alcohol response differs in heavy versus light drinkers.

Alcohol Clin Exp Res 26:827–835

Kvavilashvili L, Ellis JA (1999) The effects of positive and negative

placebos on human memory performances. Memory 7:421–437

Lepage M, Habib R, Tulving E (1998) Hippocampal PET activations

of memory encoding and retrieval: the HIPER model. Hippo-

campus 8:313–322

Lukas SE, Mendelson JH, Benedikt RA, Jones B (1986) EEG alpha

activity increases during transient episodes of ethanol-induced

euphoria. Pharmacol Biochem Behav 25:889–995

Lukas SE, Mendelson JH, Kouri E, Bolduc M, Amass L (1990)

Ethanol-induced alterations in EEG alpha activity and apparent

source of the auditory P300 evoked response potential. Alcohol

7:471–477

Marczinski CA, Fillmore MT (2005) Compensating for alcohol-

induced impairment of control: effects on inhibition and ac-

tivation of behavior. Psychopharmacology (in press)

Marlatt GA, Rohsenow D (1980) Cognitive processes in alcohol

use: expectancy and the balanced placebo design. In: Mello NK

(ed) Advances in substance abuse. J.A.I. Press, Greenwich

316

Maylor EA, Rabbitt PM, Kingstone A (1987) Effects of alcohol on

word categorization and recognition memory. Br J Psychol

78:233–239

McNair DM, Lorr M, Droppleman LF (1971) Profile of mood states.

Educational and Industrial Testing Services, San Diego

Mellanby E (1919) Alcohol: its absorption into and disappearance

from the blood under different conditions. Special report Series

no. 31. National Research Council, London

Miller ME, Adesso VJ, Fleming JP, Gino A, Lauerman R (1978)

Effects of alcohol on the storage and retrieval processes of

heavy social drinkers. J Exp Psychol Hum Learn 4:246–255

Moskowitz H, Daily J, Henderson R (1979) The Mellanby effect in

moderate and heavy drinkers. In: Johnston IR (ed) Alcohol,

drugs and traffic safety. Australian Government Publishing

Service, Canberra, pp 184–189

Mumenthaler MS, Taylor JL, O’Hara R, Yesavage JA (1999)

Gender differences in moderate drinking effects. Alcohol Res

Health 23:55–64

Nagoshi CT, Noll RT, Wood MD (1992) Alcohol expectancies and

behavioral and emotional responses to placebo versus alcohol.

Alcohol Clin Exp Res 16:255–260

Nelson TO, McSpadden M, Fromme K, Marlatt GA (1986) Effects

of alcohol intoxication on metamemory and on retrieval from

long-term memory. J Exp Psychol Gen 115:247–254

Nicholson ME, Wang M, Airhihenbuwa CO, Mahoney BS,

Christina R, Maney DW (1992) Variability in behavioral

impairment involved in the rising and falling BAC curve. J

Stud Alcohol 53:349–356

Nilsson LG, Bäckman L, Karlsson T (1989) Priming and cued recall

in elderly, alcohol intoxicated and sleep deprived subjects: a

case of functionally similar memory deficits. Psychol Med 19:

423–433

Papineau KL, Roehrs TA, Petrucelli N, Rosenthal LD, Roth T

(1998) Electrophysiological assessment (The Multiple Sleep

Latency Test) of the biphasic effects of ethanol in humans.

Alcohol Clin Exp Res 22:231–235

Parker ES, Birnbaum IM, Noble EP (1976) Alcohol and memory:

storage and state dependency. J Verbal Learn Verbal Behav

15:691–702

Parker ES, Birnbaum IM, Weingartner H, Hartley JT, Stillman RC,

Wyatt RJ (1980) Retrograde enhancement of human memory

with alcohol. Psychopharmacology 69:219–222

Parker ES, Morihisa JM, Wyatt RJ, Schwartz BL, Weingartner H,

Stillman RC (1981) The alcohol facilitation effect on memory:

a dose-response study. Psychopharmacology 74:88–92

Petersen RC (1977) Retrieval failures in alcohol state-dependent

learning. Psychopharmacology 55:141–146

Picton TW (1992) The P300 wave of the human event-related

potential. J Clin Neurophysiol 9:456–479

Pihl RO, Paylan SS, Gentes-Hawn A, Hoaken PN (2003) Alcohol

affects executive cognitive functioning differentially on the

ascending versus descending limb of the blood alcohol con-

centration curve. Alcohol Clin Exp Res 27:773–779

Pishkin V, Lawrence BE, Bourne LE Jr (1983) Cognitive and

electrophysiologic parameters during ascending and descending

limbs of the blood alcohol curve. Alcohol Clin Exp Res 7:76 –

Ray S, Bates ME, Ely BM (2004) Alcohol’s dissociation of implicit

and explicit memory processes: implications of a parallel

distributed processing model of semantic priming. Exp Clin

Psychopharmacol 12:118–125

Rimm DC, Sininger RA, Faherty JD, Whitley MD, Perl MB (1982)

A balanced placebo investigation of the effects of alcohol vs.

alcohol expectancy on simulated driving behavior. Addict Behav

7:27–32

Salinsky MC, Storzbach D, Dodrill CB, Binder LM (2001) Test-

retest bias, reliability, and regression equations for neuropsy-

chological measures repeated over a 12–16-week period. J Int

Neuropsychol Soc 7:597–605

Schacter DL (1994) Priming and multiple memory systems:

perceptual mechanisms of implicit memory. In: Schacter DL,

Tulving A (eds) Memory systems 1994. MIT Press, Cambridge,

pp 233–268

Schreckenberger M, Amberg R, Scheurich A, Lochmann M, Tichy

W, Klega A, Siessmeier T, Grunder G, Buchholz HG, Landvogt

C, Stauss J, Mann K, Bartenstein P, Urban R (2004) Acute

alcohol effects on neuronal and attentional processing: striatal

reward system and inhibitory sensory interactions under acute

ethanol challenge. Neuropsychopharmacology 29:1527–1537

Shipley WC (1940) A self-administering scale for measuring

intellectual impairment and deterioration. J Psychol 9:371–377

Sinclair JG, Lo GF (1986) Ethanol blocks tetanic and calcium-

induced long-term potentiation in the hippocampal slice. Gen

Pharmacol 17:231–233

Snodgrass JG, Feenan K (1990) Priming effects in picture fragment

completion: support for the perceptual closure hypothesis. J

Exp Psychol Gen 119:276–296

Sutker PB, Tabakoff B, Goist KC Jr, Randall CL (1983) Acute

alcohol intoxication, mood states and alcohol metabolism in

women and men. Pharmacol Biochem Behav 18:349–354

Tracy JI, Bates ME (1999) The selective effects of alcohol on

automatic and effortful memory processes. Neuropsychology

13:282–290

Tulving E (1972) Episodic and semantic memory. In: Tulving E,

Donaldson W (eds) Organization of memory. Academic, New

York, pp 381–403

Tulving E (1981) Similarity relations in recognition. J Verbal Learn

Verbal Behav 20:479–496

Tulving E (1983) Elements of episodic memory. Clarendon, Oxford

Tulving E, Watkins OC (1977) Recognition failure of words with a

single meaning. Mem Cogn 5:513–522

Tulving E, Kapur S, Craik FI, Moscovitch M, Houle S (1994)

Hemispheric encoding/retrieval asymmetry in episodic memo-

ry: positron emission tomography findings. Proc Natl Acad Sci

U S A 15:2016–2020

Wendt PE, Risberg J (2001) Ethanol reduces rCFB activation of left

dorsolateral prefrontal cortex during a verbal fluency task.

Brain Lang 77:197–215

White AM (2000) Ethanol, memory, and hippocampal function: a

review of recent findings. Hippocampus 10:88–93

Zachary RA (1986) Shipley institute of living scale, revised manual.

Western Psychological Services, Los Angeles

317