Isabel M. Birnbaum

Alcohol and memory: Retrieval processes

The influence of alcohol intoxication on the retrieval of information from memory was investigated in nonalcoholic subjects. In Experiment 1, free-recall learning of a 60-word, categorized list was impaired by alcohol intoxication. Fewer categories and words within categories were recalled by intoxicated subjects, and providing category cues after the third trial produced significantly greater improvement in recall for intoxicated than for sober subjects. It was concluded that retrieval processes may have been impaired by alcohol intoxication, but that differences in strength of the memory traces might also account for the observed differences in recall. In Experiment 2, storage was equated so that the effects of alcohol on retrieval processes alone could be assessed. Free-recall and paired-associate lists were learned in a sober State and were retrieved 1 week later in either a sober or an intoxicated state. Neither speed, nor accuracy, nor the amount of benefit produced by cues was influenced by alcohol intoxication, whereas new learning was impaired. It was concluded that the retrieval stage of memory was not affected by alcohol.

Alcohol and memory: Storage and state dependency

Abstract The effects of acute alcohol intoxication on the storage phase of memory were evaluated with two tasks that minimized response retrieval: unpaced paired-associate learning with highly available responses and forced-choice picture recognition. Paired-associate learning was impaired by the high dose of alcohol (1.0 ml/kg), while the placebo and medium dose (0.5 ml/kg) were equivalent. Picture recognition showed a dose-dependent decrement. It was concluded that storage processes are sensitive to disruption by alcohol. Alcohol did not produce decrements in memory for the original material when retention was tested 2 weeks later, and there was no evidence for state dependency.

Alcohol and elaborative schemas for sentences

Three experiments were conducted to determine whether alcohol-induced impairments of memory would be reduced by providing subjects with elaborative schemas. Anomalous sentences were presented to sober or intoxicated subjects either alone or with context words to facilitate comprehension. Memory was tested immediately with a yes-no recognition task. The results consistently showed that context words did not increase the number of hits in either sober or intoxicated subjects, and sober subjects made reliably more hits whether or not context words had been provided. The accuracy of recognition of distractors, however, was consistently improved by the presentation of context words, and was much more improved for intoxicated than for sober subjects. The results indicate that alcohol intoxication disrupted the production of elaborative schemas for understanding, but that intoxicated subjects were quite successful at utilizing such schemas as long as the schemas had been provided for them.

Study versus test trials and long-term retention in free-recall learning

A categorically structured list was learned with alternating study and test trials (Standard Group) or a single study trial followed by three successive test trials within each four-trial cycle (Recall Group). On the first test trial of each new cycle, the Standard Group and Recall Group were approximately equivalent in number of words recalled and amount of category clustering. Over a 2-day retention interval, the Standard Group and the Recall Group showed significant and equivalent losses in number of words recalled and number of categories recalled; the loss in words/category recall was of borderline significance. Either with or without cues, the Standard Group added a greater number of previously nonrecalled words than did the Recall Group during the cuing period. The results were discussed in terms of availability versus accessibility of items in different phases of the final test.

Alcohol and storage deficits: Kind of processing?

Three experiments were performed to test an explanation of alcohol-induced storage deficits that was based on the proposition that alcohol intoxication affects the kind of information-processing strategies employed by the learner. Incidental-learning procedures were used, and processing activities were manipulated through orienting task instructions. The manipulation of type of processing (phonemic or semantic), opportunity to process, and complexity of the semantic context failed to eliminate or reduce the detrimental effects of alcohol on recall performance. In all experiments, however, the pattern of results with respect to the effectiveness of encoding variables replicated the findings of numerous previous investigations (e.g., Craik & Tulving, 1975 ). The results suggest that differences in kinds of processing do not explain alcohols negative effects on memory.

Repetition, cuing, and state-dependent memory

Two experiments examined the effects of several types of repetition on state-dependent memory for conceptually categorized words. In both experiments, compatibility between pharmacological states at encoding and at retrieval facilitated the uncued recall of nonrepeated categories but had no appreciable effect on the recall of words within recalled categories. In both experiments, compatibility between encoding/retrieval states failed to facilitate the uncued recall of repeated categories li.e., categories whose names and/or exemplars were presented at least twice, in a constant or variable order, and with substantial spacing between successive presentations). And again, in both experiments, the level of uncued recall was higher for repeated than for nonrepeated categories, irrespective of the compatibility or incompatibility of encoding/retrieval states. These findings, together with other relevant observations reported in the literature, suggest that repetition in general, and repetition of category names in particular, influences recall in much the same ways as does explicit cuing with category names: It enhances the accessibility of higher order units, as reflected in uncued category recall, and it diminishes to the extent to which access to these units is state dependent.

Free-recall learning as a function of prior-list organization

The negative transfer which is observed in part-whole, free-recall learning may reflect interference from inappropriate memory units which are developed during part-list learning ( Tulving, 1966 ). In the present study, the selection of words for the part lists established varying degrees of congruence with appropriate memory units for whole-list learning. In accord with the initial hypothesis, the rate of whole-list learning was slowest when the composition of the part list prevented the development of appropriate whole-list memory units. Differences between groups in the sequence of recall of old and new items and items within the same category gave added confirmation to the hypothesis.

Is event frequency encoded automatically? The case of alcohol intoxication.

The four experiments reported show that sensitivity to event frequency is diminished by alcohol intoxication. In two experiments on confusion between presented and generated words, the function relating estimated to actual frequency was steeper for sober than intoxicated subjects. In two experiments on word-frequency estimates after a word-pronunciation task, the influence of alcohol intoxication was identified at the input as opposed to the test stage. We conclude that these findings are inconsistent with the idea that frequency-of-occurrence information is automatically encoded.

Unlearning as a function of second-list dominance

In the A-B, A-C paradigm, single interpolation (A-C) leads to less unlearning of first-list responses than does multiple interpolation (A-C, A-D, etc.) with a constant number of interpolated trials. It has been assumed that an increase in the strength of second-list responses (dominance) reduces elicitation of first-list responses during IL and thereby reduces unlearning. The present study was an attempt to manipulate the amount of unlearning in multiple-interpolation groups by varying second-list dominance prior to the introduction of new lists. Recall of first-list responses on an unpaced test was higher after single interpolation than after multiple interpolation, and the degree of second-list dominance achieved prior to the introduction of new lists had no effect on recall in the multiple-interpolation groups.

Prior-list organization in part-whole free-recall learning

In free-recall learning, the prior learning of a part list has a marked effect on the course of whole-list learning. In the present study, it was shown that S s who had learned either some or all of the words from a few of the optimal memory units on the whole list were superior in whole-list learning to S s who had learned only one item from each optimal unit. Part-list organization was quickly abandoned when it was inappropriate for whole-list learning; there was, however, persistent interference with the development of optimal memory units for whole-list learning.