Memory and Semantic Knowledge

Memory is the process of encoding, storing, and retrieving information in the brain. There are many aspects of memory. This chapter is concerned with examining the most clinically important aspects of memory, which are declarative, explicit, or conscious recollections. Patients, families, and clinicians often refer to declarative memory for recent episodic events as “short-term” memory; however, declarative memories are really long-term. Short-term (working) memory has a very different neurobiological meaning as described in this chapter. The term “amnesia,” or significant memory loss, most commonly refers to declarative, episodic memory. In addition to episodic memory, this chapter discusses another clinically important form of declarative memory, that for semantic knowledge or facts.

Memory Basics

The creation of memories is a stream that includes sensory registration, short-term/working memory, and long-term storage. The memory stream starts with sensory input and finishes with stored, accessible memory traces or engrams ( Fig. 9.1 ). After an initial brief sensory registration in the sensory organ or point of entry (auditory or echoic 10 s; visual or iconic 500 ms), there follows a period of short-term (working) memory executed in the dorsolateral frontal lobes and related areas. Short-term/working memory involves holding information online for seconds to a minute (usually 20–30 s) if unrehearsed, but much longer if the patient verbally (phonological) or visually (visuospatial) rehearses the material. Short-term/working memory is clinically difficult to differentiate from sustained attention and is usually tested as part of mental control and attentional tests such as the digit span (see Chapter 7 ). If short-term/working memory is impaired, it is difficult to proceed along the memory stream to the creation of long-term memories.

Long-term memory is the actual encoding and storage of information in the brain and can be either declarative (explicit) or nondeclarative (implicit). Declarative memory is consciously stored and retrieved information that is itself further divided into episodic (context, time, and place dependent) experiences or semantic (context, time, and place independent) facts, which are processed in the limbic and semantic areas, respectively. In comparison, nondeclarative memory includes procedural learning (motor and cognitive skills), which is mediated by the basal ganglia and cerebellum; priming, or the positive influence of a prior exposure on subsequent learning; and classical conditioning, which is facilitated by prefrontal cortex and other structures.

When talking about memory disorders and testing memory, clinicians are primarily concerned with declarative memory loss (circled in Fig. 9.1 ), often from disease affecting hippocampi or related limbic areas ( Fig. 9.2 ). For testing purposes, these long-term memories may be divisible into those that occurred either in the recent past or the remote past. Recent memory includes the capacity to remember and learn current events and information and to be able to bring it forth after an interval ranging from minutes to days, as distinguished from the inability to retrieve remote information from years past. Episodic or time-tagged recent memory difficulty is the typical memory impairment in mild cognitive impairment, dementia, or amnestic disorders. Much rarer are disorders of semantic memory for time-independent facts/knowledge. The rest of this chapter is concerned with the testing of declarative memory, both episodic and semantic. Although this chapter does not focus on testing nondeclarative memory, when necessary, clinicians can test procedural memory by the ability to learn to trace an image in a mirror and priming by the speed of recognition from word fragments of previously seen words.

Declarative Episodic Memory

MEMORY HISTORY

A complaint of memory loss can mean a great many things. “Memory loss” often describes cognitive disorders of any kind, from attentional deficits, to word-finding problems, to disturbances in executive cognition. The mental status examiner’s first step is to determine if the patient’s “memory” complaint is actually due to a disturbance in the encoding, storage, or retrieval of information. The clarification of a memory complaint involves obtaining specific examples of their memory difficulty, either from the patient or through collateral information obtained from others. For example, complaints of forgetting what they just read or what they went into a room to get suggest attentional or registration difficulties. They may indicate symptoms of anxiety or depression, physiological or sleep changes, medication or drug effects, or sensory impairment, such as hearing loss. However, complaints of forgetting what happened that morning or who they were recently introduced to suggest episodic recent memory difficulty.

The examination of memory requires patient effort and causes them anxiety and apprehension over their performance and its implications. Consequently, it is best to begin with a less threatening and indirect assessment of memory for their personal history. Ask the patient the reason and circumstances for their clinical encounter or hospitalization and how and when they arrived. The examiner may ask the patient what they think if an in-person clinic visit about current events that are salient and in the moment, for example, “what is going on in the news?” Ask for clarification of events that are sufficiently salient such that it would be reasonable for the patient to have been exposed to information on them. Difficulty relating this “incidental” information offers insight into deficits in recent memory.

RECENT MEMORY EXAMINATION

Clinicians often begin their evaluation of memory as embedded parts of mental status scales and inventories, such as the Mini-Mental State Examination or the Montreal Cognitive Assessment (see Chapters 15 and 16). These scales include a brief 3 to 5 item word list learning task. One criticism is that the interval of delayed recall and intervening distractions may not suffice for long-term memory; if there is any ongoing rehearsal, the items have not cleared from short-term/working memory. Another consideration is the inability to analyze any memory deficits by their recall processes, as described for Verbal Learning Tests.

Verbal Learning Tests. The most detailed and widely used tests of recent memory are the Verbal Learning Tests. Clinical or bedside tests are patterned after neuropsychological tests, such as the 15-item Rey Auditory-Verbal Learning Test and the 16- or 9-item versions of the California Auditory-Verbal Learning Test. A variant of these are selective reminding tests, such as the Buschke Selective Reminding Test, in which the examiner only repeatedly presents the words that the patient failed to repeat on the prior registration trial. One of the most useful Verbal Learning Tests is the “CERAD” (Consortium to Establish a Registry for Alzheimer Disease) Memory Test. It uses 10 written words, each on a different paper or card, with three registration trials, a 5-minute delayed recall interval, and a 20 word multiple-choice recognition (10 correct and 10 incorrect words).

Verbal Learning Tests of episodic recent memory assess the three recall processes of immediate recall or registration, free delayed recall, and cued recall or recognition. First, the examiner notes the patient’s registration or immediate recall of information. This process evaluates if the new information has been sufficiently attended to and held in short-term/working memory, so as to encode and store it. The ability to immediately recall and repeat back after learning trials reflects the registration process and is a prerequisite to episodic recent memory. Second, the examiner is interested in how the patient spontaneously recalls the information after a delay. Delayed recall, in which the patient must retrieve previously learned information, is the usual test for encoding and storage. Third, the examiner is interested in the retrieval of previously stored information through cuing or recognition testing. Asking the patient to recognize the previously presented items in a multiple-choice format is the usual test for retrieval. Additional processes that affect testing include visual versus verbal modality, consolidation (time needed for completion of encoding and storage), and serial processing effects (“primacy” or greater immediate recall of initial items, and “recency” or greater immediate recall of the last items).

Verbal Learning Tests involve presenting word lists for learning. The examiner can use short, bedside word list learning tasks of 3 words or up to one less than the patient’s digit span, or, preferably, a longer superspan list of 8 or 10 words ( Table 9.1 ). To prevent clustering or grouping, these word lists are composed of words that are phonemically and semantically unrelated (unless semantic clustering is of interest). As with the CERAD, the word lists can also be presented as written words, one at a time, at a regular pace in which the patient reads the word aloud. The following describes Verbal Learning Tests based on their recall processes of registration, delayed recall, and recognition.

TABLE 9.1

Verbal Learning Test: Registration and Delayed Recall

A. REGISTRATION/IMMEDIATE RECALL
To the patient: “I am going to give you a list of 10 words. Please listen to all of them, then, immediately afterward, give me as many as you can remember in any order. Ready?”
		TRIAL 1	TRIAL 2		TRIAL 3		TRIAL 4	TRIAL 5
		⬜ Cabbage	⬜ Chevrolet		⬜ Ketchup		⬜ Blue	⬜ Hockey
		⬜ Table	⬜ Blue		⬜ Juice		⬜ Rose	⬜ Juice
		⬜ Horse	⬜ Hockey		⬜ Table		⬜ Cabbage	⬜ Blue
		⬜ Chevrolet	⬜ Cabbage		⬜ Rose		⬜ Horse	⬜ Bell
		⬜ Ketchup	⬜ Horse		⬜ Bell		⬜ Hockey	⬜ Rose
		⬜ Rose	⬜ Bell		⬜ Horse		⬜ Table	⬜ Ketchup
		⬜ Bell	⬜ Rose		⬜ Cabbage		⬜ Ketchup	⬜ Chevrolet
		⬜ Blue	⬜ Table		⬜ Hockey		⬜ Chevrolet	⬜ Horse
		⬜ Juice	⬜ Juice		⬜ Blue		⬜ Bell	⬜ Table
		⬜ Hockey	⬜ Ketchup		⬜ Chevrolet		⬜ Juice	⬜ Cabbage
Correct:
False positives:
Perserverations:
Total Correct (sum of all five trials):
TOTAL NUMBER OF ERRORS:
B. DELAYED RECALL
To be administered after a 15-minute interference interval: “Try to remember the words I asked you to remember earlier. Tell me as many of the words as you can recall.”
	Remembered			Not Remembered
Cabbage						Total Number Correct:
Table
Horse
Chevrolet						Total Number of Errors:
Ketchup
Rose
Bell
Blue
Juice
Hockey