Intact Retention in Acute Alcohol Amnesia

Transcription

1 Journal of Experimental Psychology: Learning, Memory, and Cognition 1984, Vol. 10, No. 1, Copyright 1984 by the American Psychological Association, Inc. Intact Retention in Acute Alcohol Amnesia Shahin Hashtroudi George Washington University Elizabeth S. Parker National Institute on Alcohol Abuse and Alcoholism Bethesda, Maryland Lynn E. DeLisi and Richard J. Wyatt National Institute of Mental Health, Washington, DC Sharon A. Mutter George Washington University Research on alcohol amnesia has focused on memory processes that are disrupted during intoxication. The present experiment examined the possibility that certain memory processes might be resistant to the amnesic effects of alcohol. Intoxicated and sober subjects studied a list of 29 words. They were then given one of ithree different retention tests: free recall, identification of degraded words based on the procedure used by Warringtpn and Weiskrantz (1970), and yes/no recognition. As expected, free recall was significantly impaired by alcohol intoxication. In contrast, in the identification test, intoxicated subjects benefited to the same degree as sober subjects from prior exposure to the items. The two groups did not differ in immediate recognition memory. The results of the free-recall and identification tasks are similar to findings with chronic amnesic patients and suggest that perceptual fluency is not affected by alcohol, whereas elaborative processes supporting recall are particularly sensitive to disruption during intoxication. The failure to find recognition impairment at the level of intoxication used in this study distinguishes temporary alcohol amnesia from chronic amnesia. A major goal of research on alcohol and memory has been to identify the memory processes that are disrupted during intoxication (Birnbaum & Parker, 1977; Parker, Alkana, Birnbaum, Hartley, & Noble, 1974). There is compelling evidence that failure to engage in elaborative processing is a critical source of alcohol-related memory impairment (Birnbaum, Johnson, Hartley, & Taylor, 1980; Hashtroudi, Parker, DeLisi, & Wyatt, 1983). Recent studies, however, suggest that certain We are grateful to Elizabeth K. Warrington for her generous advice on the preparation of the degraded stimuli. We wish to thank Marcia K. Johnson for her comments on earlier versions of this article. Special thanks are extended to Lauri Yablick for her valuable assistance in preparing the stimulus materials and collecting the data. Beth Israel, Luanne Kallas, and Francois Lalonde ably assisted with various aspects of this research. Requests for reprints should be sent to Shahin Hashtroudi, Department of Psychology, George Washington University, Washington, DC 20052, or to Elizabeth S. Parker, who is now at the Laboratory of Psychology and Psychopathology, National Institute of Mental Health, Building 10, Room 4C-110, Bethesda, Maryland forms of memory or certain memory processes are independent of elaboration (Graf, Mandler, & Haden, 1982; Jacoby & Dallas, 1981; Warrington & Weiskrantz, 1982). These processes are relatively unimpaired in amnesic patients, including alcoholic Korsakoff patients (Jacoby, 1982; Warrington & Weiskrantz, 1982), and might also be resistant to amnesia produced by acute alcohol intoxication. The present study examines the differential vulnerability of memory processes to alcohol amnesia by drawing upon studies of the amnesic syndrome. Despite severe memory deficits in tasks requiring elaboration (Cermak, 1977; Jacoby, 1982), amnesic patients show relatively normal retention in a variety of memory tasks,. The early studies suggested that amnesics could acquire and retain motor skills (Corkin, 1968) and learn and perform maze problems (Milner, Corkin, & Teuber, 1968). In a series of classic studies, Warrington and Weiskrantz (1968, 1970) extended the domain of preserved memory processes to verbal materials. They reported that amnesics did not differ from 156

2 INTACT RETENTION IN ALCOHOL AMNESIA 157 controls in identifying visually degraded words. Neither amnesic patients nor controls had been able to identify the degraded stimuli before being exposed to these words. After the study trials, however, the groups were indistinguishable in their identification performance. More recent evidence indicates that amnesics show considerable memory savings in reading inverted texts (Cohen & Squire, 1980) and in interpreting homophones according to their recently biased meanings (Jacoby & Witherspoon, 1982). Various theoretical accounts for these results have been proposed (Cohen & Squire, 1980; Jacoby, 1982; Moscovitch, 1982; Schacter & Tulving, 1982; Warrington & Weiskrantz, 1982; Wickelgren, 1979). Jacoby (1982), for example, has suggested that due to a deficit in attention, amnesic patients are less able to engage in elaborative and effortful processing. Thus, traditional memory tests such as recall, which require elaborative processing at input and respecification of context at output, are particularly susceptible to amnesia. On the other hand, memory tests that rely on automatic processing or on perceptual fluency are resistant to amnesia. Fluency is a product of prior experience with an item and it depends on factors like repetition and perceptual similarity between study and test occurrences, Increased fluency will allow the subject to attribute a feeling of familiarity to the item (Jacoby & Dallas, 1981). According to Jacoby (1982; Jacoby & Dallas, 1981), prior experience with an item may be manifested in a variety of perceptual and classification tasks, but may not be tapped by traditional recall tasks. Identification of degraded words is an example of a task which is based on perceptual fluency (Jacoby & Dallas, 1981) but is not dependent on elaboration or interitem relationships (Jacoby, 1982; Mandler, 1980; Warrington & Weiskrantz, 1982). To examine the differential vulnerability of memory processes to alcohol amnesia, we selected three different memory tasks: free recall, identification of visually degraded words, and recognition. Intoxicated and sober subjects studied the same list of words and participated in one of three retention tests. The free-recall test, which clearly requires elaborative processing, was included as a sensitive measure of acute alcohol amnesia. Impaired recall under alcohol should, therefore, provide a comparison for possibly preserved retention during intoxication. Identification of degraded words seemed to be an ideal task for examination of preserved memory processes under alcohol. This task consisted of presentation of visually degraded words starting with the most degraded (80% degraded), and presenting increasingly more complete versions of the words (65% and 50% degraded). Subjects were asked to identify the degraded word as soon as possible. To measure the benefit from studying the items, targets were mixed with new or "distractor" items degraded in the same fashion. Inclusion of distractors was a significant deviation from the procedure used by Warrington and Weiskrantz (1970). These investigators excluded words that subjects could guess before the study trials, but they did not include distractors at test. In view of alcohol's detrimental effect on sensorimotor performance (Wallgren & Barry, 1970), it was expected that an overall deficit in identification by intoxicated subjects would occur. Comparison of performance on targets and distractors would, therefore, provide a relatively pure measure of perceptual memory. Finally, the third memory measure used here was a yes/no recognition task. Mandler (1979, 1980) has proposed that recognition can be based on familiarity, a product of intraitem integration, or on retrieval, a product of interitem organization and elaboration. Likewise, according to Jacoby (1983; Jacoby & Dallas, 1981), recognition tasks are "intermediate" between recall tasks and perceptual identification tasks, since recognition can rely on two types of information: familiarity based on relative perceptual fluency and elaboration or respecification of context. Previous studies with amnesic patients have repeatedly demonstrated impaired recognition (Huppert & Piercy, 1976; Jacoby & Witherspoon, 1982; Warrington & Weiskrantz, 1970). There are, however, surprisingly few studies on the effect of acute alcohol intoxication on recognition, and the results of these studies are mixed. Some studies (Parker, Birnbaum, & Noble, 1976; Ryback, Weinert, & Fozard, 1970) have found impaired recognition of pictures under alcohol, whereas other experiments have not (Goodwin, Powell, Bremer, Hoine, & Stern, 1969; Parker & Tulving, 1983). In

3 158 HASHTROUDI, PARKER, DELISI, WYATT, MUTTER word recognition, Wickelgren (1975) reported a small but significant impairment under alcohol in a yes/no continuous recognition test, but others failed to find a detrimental effect of alcohol (Goodwin, Hill, Powell, & Viamontes, 1973; Taylor, Johnson, & Birnbaum, 1980). Thus, in view of these unexplained mixed results on the effect of alcohol on recognition, it seemed important to include a recognition test in the present study. > Design and Materials Method The design was a 2 X 3 factorial with state (sober vs. intoxicated) and type of retention condition (free recall, identification of degraded words, or recognition) as between-subject variables. Sixteen subjects were tested in each group. In accordance with the procedure used by Warrington and Weiskrantz (1970), potential target and distractor items were selected from the Thorndike and Lorge (1944) word count Two hundred and forty-eight items were chosen from the total set of five-letter words of various grammatical classes, with frequencies of either AA, A, or occurrences per million. An attempt was made to select only those items that were visually distinctive. Words with obvious semantic similarity were eliminated. Examples of degraded words are shown in Figure 1. Visual degradation was accomplished by filtering the original items (Warrington & Weiskrantz, 1968, 1970). The area covered by a word was divided into a grid, and random portions of the grid pattern were blocked to create a filter. The three filters that were made by this method obscured either 50%, 65%, of 80% of the area. Slides for the visually degraded words were obtained by photographing a word through each of these filters. Warrington and Weiskrantz (1970) used a preliminary task to exclude any word that could be identified by a subject at the 80% degradation level. Thus, each list for each individual subject comprised words that could not be identified by the subject. In the present experiment, identifiabiiity of words was determined by a pilot experiment. The critical pool of 248 items at their most degraded level (80%) was shown to 7 pilot subjects selected from the same population as the experimental subjects. If 4 or more subjects were able to guess the identity of a word, it was eliminated from the set to be used for later testing. A total of 132 items were eliminated by this procedure. The remaining words were divided into four 29-item word lists matched for frequency of occurrence and identification characteristics of the items. All four lists comprised approximately the same number of AA-A words (21 or 20) and words (8 or 9). Likewise, all lists contained about the same number of items identified by 3, 2, 1, or 0 subjects during the pilot testing. Two of these lists (1 and 2) were used as study lists. Half of the subjects in each test condition received List 1 at acquisition and the remaining half received List 2. The other two lists (3 and 4) were used as new or distractor items in test lists for the identification and recognition conditions. List 3' new items FORCE OLIVE WIDOW Figure 1. Examples of degraded words. were combined with List 1 study items, and List 4 new items were combined with List 2 study items. There were two random orders of each list used for study and two random orders of each test list (targets and distractors combined). Each random order of the study lists was combined equally often with each random order of the test lists. The words were arranged randomly in both test lists with two restrictions: (a) The words from the beginning and the end of the study lists were never placed first or last on the test lists, and (b) no more than two adjacent words from the study list were presented successively in the test lists. All study and test items were presented using a Kodak Carousel slide projector with a built-in timing mechanism and a detachable timer. The projection screen was covered with cloth to reduce glare. The viewing distance from the seated subject was approximately 6 ft (1.83 m). Subjects ; Ninety-six male volunteers between the ages of 21 and 35 were recruited by advertisements in the George Washington University and University of Maryland newspapers; they were paid $50 for 1 full day of participation in the experiment and an additional $5 for; every hour beyond the initial 8 hours. Volunteers were screened initially by a brief telephone interview. Only subjects who reported consumption of five drinks in a single session during the past 6 months, the dose equivalent to the one used in the experiment, were scheduled for full screening. Each subject was screened by a physician. The screening included a physical examination, medical history, alcohol and drug history, psychiatric evaluation, blood, chemistry and hematological evaluation. Subjects who showed evidence of physical or psychiatric disorders were excluded. Each subject gave written informed consent before his participation. The subjects were instructed to refrain from the use of drugs (except coffee and cigarettes);and alcohol for 48 hours prior to participation in the study, and they were asked to fast from midnight until the time of arrival at the experiment. Each subject who received alcohol remained in the laboratory until his blood alcohol concentration returned to zero, approximately 7 hours after he began to drink. Alcohol Procedures Subjects arrived at the laboratory at 8:15 a.m. and were weighed and interviewed about their use of drugs during

4 the preceding 48 hours; a baseline blood alcohol concentration of zero was ascertained. Blood alcohol concentration in this experiment was measured by breath analysis on a Mark IV Gas Chromatograph Intoximeter. Subjects were randomly assigned to one of the six experimental conditions. Following preliminary instructions and prior to alcohol or placebo administration, all subjects received a package of Carnation Instant Breakfast with 4 oz ( ml) of skim milk and 4 oz ( ml) of water. They then either watched television or read magazines until called for the experiment 1 hour later. Alcohol and placebo drinks were administered according to procedures designed to mask the" alcohol content of the beverages (Parker etal., 1976). Exactly 1 hour after breakfast, all subjects received four drinks, and each drink was consumed steadily over a 10-min interval. For the subjects in the intoxicated groups, each drink contained 0.25 ml/ kg ethanol with one part 95% ethanol and five parts masking solution (Minute Maid Lemonade Crystals 150% recommended concentration and 1.0 ml peppermint extract per liter solution). Thus, subjects in the intoxicated groups consumed 1.0 ml ethanol per kilogram of body -weight over 40 min. The subjects in the placebo condition also received four drinks over a period of 40 min. The placebo drinks were prepared with six parts masking solution and the rims of glasses were swabbed with ethanol. After a 20- min absorption period, during which the subjects were engaged in conversation with the experimenter, a breath reading was taken and the experimental task began. Procedure All subjects were tested individually and were given general instructions about the memory test before ingestion of their drinks. After the drinking and absorption period, subjects were told that they would see a list of words presented one at a time on the projection screen. They were instructed that their memory for the items would be tested but the type, of memory test was not specified, The 29 words were then presented in nondegraded fashion on a single trial. All words were presented at the rate of 2 s per item with an interval of 1 s between items. To ensure attention, subjects read each word aloud as it was presented. Three buffer words preceded and followed the presentation of the study word list. These buffer words were not tested in the recognition and identification conditions and were excluded from analyses in the recall condition. Following acquisition, subjects performed an arithmetic distractor task involving addition, subtraction, and multiplication problems. The distractor task lasted for 5 min, after which different groups of subjects performed one of the three retention tests: identification, yes/no recognition, or free recall. Subjects in the identification group were shown the degraded versions of 29 old study items mixed with 29 new test items. They were told that some of the words represented by the degraded stimuli had been presented on the study list and others had not. The goal was to identify the words as soon as possible. Subjects were first shown the most degraded version (80%) of an item. If the item was not identified at this stage, the 65% degraded version was presented. If the item still remained unidentified, the least degraded (50%) version was shown. The later versions were not presented if the subject identified an item at an INTACT RETENTION IN ALCOHOL AMNESIA 159 earlier stage. An experimenter recorded responses.and identified the word if the subject failed to do so at the 50% level. Each degraded word remained on the screen for 4 s with a 1-s interval between words. In the recognition condition 29 study words were shown with 29 distractors. The complete word was, of course, used here. Each word remained on the screen for 7 s with a 1-s interval between words. During this time, subjects recorded their recognition decisions on sheets of- paper containing blank spaces for a yes or no judgment. In addition, they circled a number on a 4-point rating scale indicating their confidence in their decision. Subjects in the recall group were given 5 min to write down as many words as they could. A breath alcohol reading was taken after completion of each test. Results Blood Alcohol Concentration The mean blood alcohol level (BAG) for the intoxicated subjects was.08 g/100 ml before and after the experiment. The BAC was zero for the sober subjects. Retention Data The proportions of responses in the three test conditions are shown in Table 1. The corrected recognition score is the difference-between the proportion of hits and false alarms. For the identification task, the proportion of old words is the number of old (study) words that a subject identified in the most degraded form (80%) divided by the total number of old words (29). Similarly, the proportion of new words is the number of new items identified at the 80% level divided by the total number of new words (29).The corrected score is the difference in proportions of old and new words identified. Thus, the proportion of new items identified provides a baseline measure of perceptual identification proficiency, and the corrected score provides a relatively pure measure of improvement from having studied the words (i.e., a measure of memory). All statistical tests were conducted at a p <.05 level unless otherwise specified. The effect of alcohol was clearly dependent on how memory was tested (see Table 1). A 3 X 2 analysis of variance (ANOVA) on the proportion of correctly recalled items and the corrected scores from the identification and recognition conditions revealed a significant effect of retention condition, F(2,90) = 329:88, MS; =.007, a significant effect of state, F(l, 90) = 490, and a significant interaction of

5 160 HASHTROUDI, PARKER, DELISI, WYATT, MUTTER Table 1 Proportions of responses in the test conditions Measure Recall Identification Old' New" Corrected Recognition Old New c Corrected d«sober State Intoxicated " Number of old words identified at the 80% level divided by total number of old words (29). b Number of new words identified at the 80% level divided by total number of new words (29). 'Proportion of false alarms. "Average of d' values calculated for individual subjects. state and retention condition, F(2,90) = Analyses of simple effects determined that the interaction was due to a large effect of state in the recall condition, F(l, 90) = 18.14, MS e =.007. The free-recall condition was clearly sensitive to impairment by alcohol and did, therefore, provide an index of amnesia for comparison with the other test conditions. There were no state-related differences in the identification, F(l, 90) = 0.57, MS e =.007, and recognition conditions, F(l, 90) = 0.14, MS^= In view of the severe impairment in free recall, the condition that depends on elaborative processing, the absence of impairment on identification or recognition is particularly striking. Further analyses were conducted to examine these results more closely. An ANOVA on the proportion of correctly identified old and new items at the 80% degradation level (see Table 1) showed that both intoxicated and sober subjects identified more old items than new items, F(l, 30) = 53.83, MS e =.003, and intoxicated subjects identified fewer items (both old and new) than sober subjects, F(l, 30) = 14.76, MS e =.015. The interaction of state and type of item, however, was not significant, F(l, 30) = 1.00, MS e =.003. Thus, intoxicated subjects had some impairment in identification of degraded words, but they benefited to the same degree as sober subjects from a single exposure to the item at study. It should be noted that the different baseline identification levels for intoxicated and sober subjects (12% vs. 25%) present some difficulty for assessing the improvement from prior exposure to the items. For example, in terms of percentage gain, intoxicated subjects benefited more from prior presentation of the items than sober subjects (92% vs. 36%). However, the various measures of improvement (corrected identification score, percentage gain, and the absence of an interaction between state and type of item) consistently indicated that intoxicated subjects were not impaired in retention as assessed by identification of degraded words. To explore the identification performance further, an ANOVA was performed on the mean level of identification of old items and new items (80%, 65%, and 50%) for both groups. Mean level of identification was determined by assigning a score of 3 if subjects identified the item at the 80% level, and scores of 2 and 1 if the subjects identified the items at the 65% and 50% levels, respectively. A score of 0 was assigned if subjects failed to identify the word at the 50%level. Mean levels of identification for the old and new items were 2.11 and 1.91 for sober subjects, showing an improvement of 0.20 from prior study of qld items. The corresponding scores for old and new items for intoxicated subjects were 1.85 and 1.66, with an improvement of Once again, both sober and intoxicated subjects identified the old items at a higher level of degradation than new items, F(l, 30) = 16.94, MS e = 35, and the sober subjects were more likely to identify both old and new items at a higher degradation level than were the intoxicated subjects, F(l, 30) = 8.90, MS f =.119. The interaction of state by type of item, however, was not significant, F(\, 30) < 1, MS e =.035. This analysis, therefore, confirms that previous exposure to an item facilitates later identification of that item, and that this facilitation is not different for sober and intoxicated subjects. Further analyses of recognition data presented in Table 1 showed that intoxicated subjects had a lower hit rate, r(30) = 1.91, p <.05, one-tailed, and a marginally lower false positive rate, «(30) = 1.31, p <, 10, one-tailed, than sober subjects. Nevertheless,, in agreement with the previously reported corrected rec-

6 ognition scores, alcohol did not have a significant effect on d' values (see Table 1), f(30) = Overall, it appears that the particular form of immediate recognition used in the present experiment is spared in alcohol amnesia. Discussion INTACT RETENTION IN ALCOHOL AMNESIA 161 The results of the recall and identification conditions suggest that memory tasks requiring elaboration or interitem integration are particularly susceptible to alcohol amnesia, whereas tasks relying on perceptual fluency are resistant to disruption by alcohol. These results suggest certain similarities between temporary alcohol-induced amnesia and chronic amnesias produced by various sources (Warrington & Weiskrantz, 1970, 1982). This similarity, however, does not extend to the recognition results. 1 In agreement with previous research (e.g., Birnbaum & Parker, 1977), alcohol induced a significant decrement in free recall. A freerecall test, of course, requires interitem integration and elaboration. Since other studies have shown that elaborative processes are disrupted under alcohol with the same dose used in this experiment (Birnbaum et al., 1980; Hashtroudi et al., 1983), it seems reasonable to conclude that the observed recall deficit reflects a disruption in the formation of interitem relations and elaborative processing during intoxication. In contrast to the marked deficit in recall, retention as assessed by identification of degraded words was not impaired. Intoxicated subjects benefited to the same degree as sober controls from a single exposure to the item at study. The identification task has been characterized as a test of perceptual memory (Jacoby, 1982) or "pure" familiarity (Mandler, 1980), Regardless of the specific characterization, the identification task seems to be independent of elaborative processing (Jacoby, 1982; Mandler, 1980; Warrington & Weiskrantz, 1982). Preserved identification performance under alcohol suggests that alcohol has differential effects on perceptually based versus elaboratively based memory processes. The dissociation between impaired recall and unimpaired fragment identification during alcohol amnesia is similar to the results obtained with amnesic patients (Warrington & Weiskrantz, 1970). Impairment in free recall has consistently been demonstrated for a variety of amnesic disorders (Warrington & Weiskrantz, 1970). In fact, the magnitude of recall impairment in the present study is not unlike the magnitude of deficit reported by Warrington and Weiskrantz (1970) for amnesic patients. In this study, recall performance for intoxicated subjects was 43% of performance for sober subjects (21% recall for sober subjects and 9% recall for intoxicated subjects), whereas amnesic patients' recall level was 61% of control subjects' performance (54% recall for controls and 33% recall for amnesic patients with shorter lists than those used here). Similarly, preserved identification performance has been repeatedly demonstrated with amnesic patients (Warrington & Weiskrantz, 1968,1970) and in patients who received electroconvulsive therapy (Squire, Wetzel, & Slater, 1978). The present findings are generally consistent with these studies: The benefit from prior presentation of the items was the same for sober and intoxicated subjects, although the absolute level of identification was lower for intoxicated subjects. One exception to the parallel between acute alcohol amnesia and chronic amnesia is the performance on the recognition test. Recognition memory is impaired in chronic amnesics (Cohen & Squire, 1980; Huppert & Piercy, 1976; Jacoby & Witherspoon, 1982; Piercy, 1977; Warrington & Weiskrantz, 1970). In fact, impaired recognition is considered to be a criterion for diagnosis of the amnesic syndrome by certain investigators (Warrington & Weiskrantz, 1982). The particular form of recognition test used in the present study, however, was not impaired by alcohol. Intoxicated and sober subjects did not differ in d' values and corrected recognition scores. 1 Our conclusions regarding the effect of alcohol on various tasks are limited to the dose used here (equivalent to five or six drinks) and do not generalize to higher levels of intoxication. It should be noted, however, that this dose is commonly used in research on alcohol intoxication (cf. Birnbaum et al., 1980; Parker et al., 1976), and it is the highest dose that can be administered in a relatively brief period of time (40 min) without undesirable side effects. For several reasons, higher levels of intoxication are difficult to investigate with nonalcoholic subjects.

7 162 HASHTROUDI, PARKER, DELISI, WYATT, MUTTER Moreover, as noted earlier, alcohol-induced decrements in recognition are not always demonstrated (Goodwin et al., 1969, 1973; Parker & TAilving, 1983; Taylor et al., 1980), and when they appear they seem to be of relatively small size (Wickelgren, 1975). It is not clear why the level of intoxication that disrupts recall does not consistently impair recognition. It is possible, however, that the disparate results arise from the different types of information that can be used for a recognition decision. It is often assumed that both familiarity (derived primarily from perceptual characteristics of the item) and elaboration are required for a recognition decision (Jacoby & Dallas, 1981; Mandler, 1980). It can be argued, however, that the demand for elaboralive processing at retrieval may vary from one recognition test to another. The detrimental effect of alcohol may occur when recognition judgments are heavily dependent on elaborative processing and respeciffcation of context. This account, however, is speculative and there is no direct evidence to support it. The small deficit in recognition under alcohol and the failure to find a recognition impairment in some experiments contrast with the severe recognition impairment in amnesic patients. Other differences in performance of intoxicated subjects and amnesic patients, particularly alcoholic Korsakoff patients, also suggest that the amnesia resulting from the level of acute alcohol intoxication used here and the memory deficit in chronic amnesic patients are not identical. Korsakoff patients show an abnormally high false-alarm rate in yes/no recognition tests (Huppert & Piercy, 1976), whereas intoxicated subjects do not. The high false-alarm rate in Korsakoff patients could be attributed to the increased susceptibility to interference or the perseveration that is characteristic of these patients (Moscovitch, 1982; Squire, 1982). Moreover, the nature and the degree of amnesia in Korsakoff patients is so severe that immediately after completion of a test, they might deny any familiarity with the task itself. This striking separation between objective evidence of learning and conscious memory for the task does not occur with intoxicated college students in a laboratory situation. It is important to identify additional aspects of memory functioning that differentiate between acute and chronic alcohol amnesia. However, comparisons across amnesias of different etiologies with different samples of subjects should be made with great caution (see Squire, 1982). Overall, the present results on alcohol-induced amnesia support the distinction between two forms of memory, a distinction already suggested by studies of the amnesic syndrome and normal memory (Cohen & Squire, 1980; Graf et al., 1982; Jacoby & Dallas, 1981; Jacoby & Witherspoon, 1982; Moscovitch, 1982; Schacter & Tulving, 1982; Tulving, Schacter, & Stark, 1982; Warrington & Weiskrantz, 1982). We have offered only one account of the dissociation phenomena, that is, a dissociation between perceptually based and elaboratively based memory processes (Graf et al., 1982; Jacoby & Dallas, 1981). Others have suggested a dissociation between semantic memory and memory processes requiring cognitive mediational strategies (Warrington & Weiskrantz, 1982), a dissociation between skill or procedural memory and a cognitive conscious memory system (Cohen & Squire, 1980; Moscovitch, 1982), or a dissociation between episodic memory and a;third memory system that underlies the persistent effect of prior experience on fragment completion (Schacter & Tulving, 1982; Tulving et al., 1982). Clearly our knowledge of the underlying memory processes responsible for dissociation phenomena is far from complete. Nevertheless, the distinction.between different forms of memory necessitates a new approach for investigating the effect of alcohol on memory. This approach entails an examination of what is, as well as what is not, remembered during intoxication. References Birnbaum, I. M., Johnson, M. K., Hartley, J. X, & Taylor, T. H. (1980). Alcohol and elaborative schemes for sentences. Journal of Experimental Psychology: Human Learning and Memory, 6, Birnbaum, I. M., & Parker, E. S. (1977), Acute effects of alcohol on storage and retrieval. In I. M. Birnbaum & E. S. Parker (Eds.), Alcohol and human memory (pp ). Hfflsdale, NJ: Erlbaum. Cermak, L. S. (1977). The contribution of a "processing" deficit to alcoholic Korsakoff patients' memory disorder. In I. M. Birnbaum & E. S. Parker (Eds.), Alcohol and human memory (pp ). Hillsdale, NJ: Erlbaum. Cohen, N. J., & Squire, L. R. (1980).;Preserved learning and retention of pattern-analyzing skill in amnesia: Dis-

8 INTACT RETENTION IN ALCOHOL AMNESIA 163 sociation of knowing how and knowing that. Science, 210, Corkin, S. (1968). Acquisition of motor skill after bilateral medial temporal-lobe excision. Neuropsychologia, 6, Goodwin, D; W., Hill, S. Y., Powell, B., & Viamontes, J. (1973). Effect of alcohol on short-term memory in alcoholics. British Journal of Psychiatry, 122, Goodwin, D. W., Powell, B., Bremer, D., Hoine, H,, & Stern, J. (1969). Alcohol and recall: State-dependent effects in man. Science, 163, Graf, P., Mandler, G., & Haden, P. E. (1982). Simulating amnesic symptoms in normal subjects. Science, 218, , Hashtroudi, S., Parker, E. S., DeLisi, L. E., & Wyatt, R. J, (1983). On elaboration and alcohol. Journal of Verbal Learning and Verbal Behavior, 22, Huppert, F. A., & Piercy, M. (1976). Recognition memory in amnesic patients: Effect of temporal context and familiarity of material. Cortex, 12, Jacoby, L. L. (1982). Knowing and remembering: Some parallels in the behavior of Korsakoff patients and normals. In L. S. Cermak (Ed,), Human memory and amnesia (pp ). Hillsdale, NJ: Erlbaum, Jacoby, L. L, (1983). Perceptual enhancement: Persistent effects of an experience. Journal of Experimental Psychology: Learning, Memory, and Cognition, 9, Jacoby, L. L., & Pallas, M. (1981). On the relationship between autobiographical memory and perceptual learning. Journal of Experimental Psychology: General, 110, Jacoby, L. L., & Witherspoon, D. (1982). Remembering without awareness. Canadian Journal of Psychology, 36, Mandler, G. (1979). Organization and repetition: Organizational principles with special reference to rote learning, In L. Nilsson (Ed.), Perspectives in memory research (pp ). Hillsdale, NJ: Erlbaum. Mandler, G. (1980). Recognizing: The judgment of previous occurrence. Psychological Review, 87, Milner, B., Corkin, S., & Teuber, H. L. (1968). Further analysis of the hippocampal amnesic syndrome: 14- year follow-up study of H.M. Neuropsychologia, 6, Moscovitch, M. (1982). Multiple dissociations of function in amnesia. In L. S, Cermak (Ed,), Human memory and amnesia (pp ). Hillsdale, NJ: Erlbaum. Parker, E. S., Alkana, R. L., Birnbaum, I. M., Hartley, J. T., & Noble, E. P. (1974). Alcohol and the disruption of cognitive processes. Archives of General Psychiatry, 31, Parker, E. S,, Birnbaum, I. M., & Noble, E. P. (1976). Alcohol and memory: Storage and state-dependency. Journal of Verbal Learning and Verbal Behavior, 15, Parker, E. S., & Tulving, E. (1983), [Picture recognition on the ascending and descending limbs of the blood alcohol curve]. Unpublished raw data. Piercy, M. (1977). Experimental studies of the organic amnesic syndrome. In C. W, M. Whitty & O. L. Zangwill (Eds.), Amnesia (pp. 1-51). London: Butterworths. Ryback, R. S., Weinert, J., & Fbzard, J. L. (1970). Disruption of short-term memory in man following consumption of ethanol. Psychonomic Science, 20, Schacter, D. L., & Tulving, E. (1982). Memory, amnesia, and the episodic/semantic distinction. In R. L. Isaacson & N. E. Spear (Eds.), The expression of knowledge (pp ). New York: Plenum Press. Squire, L. R. (1982). Comparisons between forms of amnesia: Some deficits are unique to Korsakoff's syndrome. Journal of Experimental Psychology: Learning, Memory, and Cognition, 8, Squire, L. R., Wetzel, C. D., & Slater, P. C. (1978). Anterograde amnesia following ECT: An analysis of beneficial effects of partial information. Neuropsychologia, M Taylor, T.H., Johnson, M. K., & Birnbaum, I. M. (1980). Alcohol intoxication and reality monitoring. Unpublished manuscript. Thorndike, E. L., & Lorge, I. (1944). The teachers' word book of 30,000 words, New iork: Columbia University. Tulving, E, Sehacter, D. L., & Stark, H. A. (1982). Priming effects in word-fragment completion are independent ofrecognition memory. Journal of Experimental Psychology: Learning, Memory, andcognition, 8, Wallgren, H., & Barry, H. (1970). Actions of alcohol: Biochemical, physiological and psychological aspects. Amsterdam: Elsevier. Warrington, E. K., & Weiskrantz, L. (1968). New method of testing long-term retention with special reference to amnesic patients. Nature, 217, Warrmgton, E. K., & Weiskrantz* L. (1970). Amnesic syndrome: Consolidation or retrieval? Nature, 228, Warrington, E. K., & Weiskrantz, L. (1982). Amnesia; A disconnection syndrome? Neuropsychologia, 20, Wickelgren, W. A. (1975). Alcoholic intoxication and memory storage dynamics. Memory & Cognition, 3, Wickelgren, W. A, (1979). Chunking and consolidation: A theoretical synthesis of semantic networks, configuring in conditioning, S-R versus cognitive learning, normal forgetting, the amnesic syndrome and the hippocampal arousal system. Psychological Review, 86, 44-60, Received September 29, 1982 Revision received June 22, 1983