Econstudentlog

Principles of memory (II)

I have added a few more quotes from the book below:

Watkins and Watkins (1975, p. 443) noted that cue overload is “emerging as a general principle of memory” and defined it as follows: “The efficiency of a functional retrieval cue in effecting recall of an item declines as the number of items it subsumes increases.” As an analogy, think of a person’s name as a cue. If you know only one person named Katherine, the name by itself is an excellent cue when asked how Katherine is doing. However, if you also know Cathryn, Catherine, and Kathryn, then it is less useful in specifying which person is the focus of the question. More formally, a number of studies have shown experimentally that memory performance systematically decreases as the number of items associated with a particular retrieval cue increases […] In many situations, a decrease in memory performance can be attributed to cue overload. This may not be the ultimate explanation, as cue overload itself needs an explanation, but it does serve to link a variety of otherwise disparate findings together.”

Memory, like all other cognitive processes, is inherently constructive. Information from encoding and cues from retrieval, as well as generic information, are all exploited to construct a response to a cue. Work in several areas has long established that people will use whatever information is available to help reconstruct or build up a coherent memory of a story or an event […]. However, although these strategies can lead to successful and accurate remembering in some circumstances, the same processes can lead to distortion or even confabulation in others […]. There are a great many studies demonstrating the constructive and reconstructive nature of memory, and the literature is quite well known. […] it is clear that recall of events is deeply influenced by a tendency to reconstruct them using whatever information is relevant and to repair holes or fill in the gaps that are present in memory with likely substitutes. […] Given that memory is a reconstructive process, it should not be surprising to find that there is a large literature showing that people have difficulty distinguishing between memories of events that happened and memories of events that did not happen […]. In a typical reality monitoring experiment […], subjects are shown pictures of common objects. Every so often, instead of a picture, the subjects are shown the name of an object and are asked to create a mental image of the object. The test involves presenting a list of object names, and the subject is asked to judge whether they saw the item (i.e., judge the memory as “real”) or whether they saw the name of the object and only imagined seeing it (i.e., judge the memory as “imagined”). People are more likely to judge imagined events as real than real events as imagined. The likelihood that a memory will be judged as real rather than imagined depends upon the vividness of the memory in terms of its sensory quality, detail, plausibility, and coherence […]. What this means is that there is not a firm line between memories for real and imagined events: if an imagined event has enough qualitative features of a real event it is likely to be judged as real.”

“One hallmark of reconstructive processes is that in many circumstances they aid in memory retrieval because they rely on regularities in the world. If we know what usually happens in a given circumstance, we can use that information to fill in gaps that may be present in our memory for that episode. This will lead to a facilitation effect in some cases but will lead to errors in cases in which the most probable response is not the correct one. However, if we take this standpoint, we must predict that the errors that are made when using reconstructive processes will not be random; in fact, they will display a bias toward the most likely event. This sort of mechanism has been demonstrated many times in studies of schema-based representations […], and language production errors […] but less so in immediate recall. […] Each time an event is recalled, the memory is slightly different. Because of the interaction between encoding and retrieval, and because of the variations that occur between two different retrieval attempts, the resulting memories will always differ, even if only slightly.”

In this chapter we discuss the idea that a task or a process can be a “pure” measure of memory, without contamination from other hypothetical memory stores or structures, and without contributions from other processes. Our impurity principle states that tasks and processes are not pure, and therefore one cannot separate out the contributions of different memory stores by using tasks thought to tap only one system; one cannot count on subjects using only one process for a particular task […]. Our principle follows from previous arguments articulated by Kolers and Roediger (1984) and Crowder (1993), among others, that because every event recruits slightly different encoding and retrieval processes, there is no such thing as “pure” memory. […] The fundamental issue is the extent to which one can determine the contribution of a particular memory system or structure or process to performance on a particular memory task. There are numerous ways of assessing memory, and many different ways of classifying tasks. […] For example, if you are given a word fragment and asked to complete it with the first word that pops in your head, you are free to try a variety of strategies. […] Very different types of processing can be used by subjects even when given the same type of test or cue. People will use any and all processes to help them answer a question.”

“A free recall test typically provides little environmental support. A list of items is presented, and the subject is asked to recall which items were on the list. […] The experimenter simply says, “Recall the words that were on the list,” […] A typical recognition test provides more environmental support. Although a comparable list of items might have been presented, and although the subject is asked again about memory for an item in context, the subject is provided with a more specific cue, and knows exactly how many items to respond to. Some tests, such as word fragment completion and general knowledge questions, offer more environmental support. These tests provide more targeted cues, and often the cues are unique […] One common processing distinction involves the aspects of the stimulus that are focused on or are salient at encoding and retrieval: Subjects can focus more on an item’s physical appearance (data driven processing) or on an item’s meaning (conceptually driven processing […]). In general, performance on tasks such as free recall that offer little environmental support is better if the rememberer uses conceptual rather than perceptual processing at encoding. Although there is perceptual information available at encoding, there is no perceptual information provided at test so data-driven processes tend not to be appropriate. Typical recognition and cued-recall tests provide more specific cues, and as such, data-driven processing becomes more appropriate, but these tasks still require the subject to discriminate which items were presented in a particular specific context; this is often better accomplished using conceptually driven processing. […] In addition to distinctions between data driven and conceptually driven processing, another common distinction is between an automatic retrieval process, which is usually referred to as familiarity, and a nonautomatic process, usually called recollection […]. Additional distinctions abound. Our point is that very different types of processing can be used by subjects on a particular task, and that tasks can differ from one another on a variety of different dimensions. In short, people can potentially use almost any combination of processes on any particular task.”

Immediate serial recall is basically synonymous with memory span. In one the first reviews of this topic, Blankenship (1938, p. 2) noted that “memory span refers to the ability of an individual to reproduce immediately, after one presentation, a series of discrete stimuli in their original order.”3 The primary use of memory span was not so much to measure the capacity of a short-term memory system, but rather as a measure of intellectual abilities […]. Early on, however, it was recognized that memory span, whatever it was, varied as function of a large number of variables […], and could even be increased substantially by practice […]. Nonetheless, memory span became increasingly seen as a measure of the capacity of a short-term memory system that was distinct from long-term memory. Generally, most individuals can recall about 7 ± 2 items (Miller, 1956) or the number of items that can be pronounced in about 2 s (Baddeley, 1986) without making any mistakes. Does immediate serial recall (or memory span) measure the capacity of short-term (or working) memory? The currently available evidence suggests that it does not. […] The main difficulty in attempting to construct a “pure” measure of immediate memory capacity is that […] the influence of previously acquired knowledge is impossible to avoid. There are numerous contributions of long-term knowledge not only to memory span and immediate serial recall […] but to other short-term tasks as well […] Our impurity principle predicts that when distinctions are made between types of processing (e.g., conceptually driven versus data driven; familiarity versus recollection; automatic versus conceptual; item specific versus relational), each of those individual processes will not be pure measures of memory.”

“Over the past 20 years great strides have been made in noninvasive techniques for measuring brain activity. In particular, PET and fMRI studies have allowed us to obtain an on-line glimpse into the hemodynamic changes that occur in the brain as stimuli are being processed, memorized, manipulated, and recalled. However, many of these studies rely on subtractive logic that explicitly assumes that (a) there are different brain areas (structures) subserving different cognitive processes and (b) we can subtract out background or baseline activity and determine which areas are responsible for performing a particular task (or process) by itself. There have been some serious challenges to these underlying assumptions […]. A basic assumption is that there is some baseline activation that is present all of the time and that the baseline is built upon by adding more activation. Thus, when the baseline is subtracted out, what is left is a relatively pure measure of the brain areas that are active in completing the higher-level task. One assumption of this method is that adding a second component to the task does not affect the simple task. However, this assumption does not always hold true. […] Even if the additive factors logic were correct, these studies often assume that a task is a pure measure of one process or another. […] Again, the point is that humans will utilize whatever resources they can recruit in order to perform a task. Individuals using different retrieval strategies (e.g., visualization, verbalization, lax or strict decision criteria, etc.) show very different patterns of brain activation even when performing the same memory task (Miller & Van Horn, 2007). This makes it extremely dangerous to assume that any task is made up of purely one process. Even though many researchers involved in neuroimaging do not make task purity assumptions, these examples “illustrate the widespread practice in functional neuroimaging of interpreting activations only in terms of the particular cognitive function being investigated (Cabeza et al., 2003, p. 390).” […] We do not mean to suggest that these studies have no value — they clearly do add to our knowledge of how cognitive functioning works — but, instead, would like to urge more caution in the interpretation of localization studies, which are sometimes taken as showing that an activated area is where some unique process takes place.”

October 6, 2018 - Posted by | Biology, Books, Psychology

No comments yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: