Remembering to do things in the laboratory and everyday life

Epilepsy can give rise to memory problems that are transient, long-standing or more-or-less permanent, depending on epileptic discharges, drug therapy, or brain lesions. The memory deficits are therefore potentially very complex and I shall not attempt to separate the various sources of deficit. Instead I shall take an entirely pragmatic perspective, with little in the way of intervention from theory. I shall begin by describing various techniques for the investigation of the failures of memory that occur in everyday life, the strengths and weaknesses of these techniques, and how they can be improved. I shall argue that we need to adapt the way in which memory is conventionally examined in the clinical context so as to take account of the limitations inherent in the psychometric tradition from which much of the work in this area stems. We need to tackle the issues regarding the validity of the tests we use, above all their validity in everyday life. The techniques for measuring memory deficits can be divided into those that seek to describe memory in the context of everyday life, and those that measure memory in the laboratory or examining room. The former techniques obviously have a greater potential relevance to everyday life a greater "ecological validity", to use the overused jargon. The latter techniques have a greater potential reliability, partly because the examining room provides few of the highly variable and idiosyncratic cues that occur in everyday life. Measures that are unreliable cannot be valid so there may well be a trade-off: attempts to increase ecological validity may actually decrease predictive validity by virtue of a decrease in reliability. So how well do those techniques that have attempted to measure memory in everyday life fare? Just how valid are they? There are several techniques that have been described in the literature. Most use an observer to note down those aspects of everyday behaviour that suggest some failure of memory. Sometimes the observer may be required to fill out a questionnaire that asks how often certain incidents occur. For example, the observer might be asked to judge how frequently the patient forgets where he has put something, how often he loses things around the house, or to judge how frequently he fails to recognise places that he has been to before. Instead of simply making a judgement as to how often these events occur, the observer may be asked to keep a diary, and at the end of each day to make a note of any failures of memory. Diaries with checklists have the advantage of recording actual events but they are, of course, relatively time-consuming. Sometimes the patient himself may be the observer, sometimes a relative may take this role. When the patient is the observer it is possible to record those failures of memory that do not give rise to observable behaviour. On the other hand when a relative of the patient does the observation the memory failures may perhaps be less likely to be forgotten before they are recorded. The above advantages and disadvantages of the various techniques were borne out in a study by Sunderland and co-workers (1) who investigated memory after severe head injury. They addressed the issue of the validity of the techniques in a variety of ways. For example the intercorrelations between the various techniques were reasonably high, although generally lower for questionnaires than for diary checklists, and lower when the patient himself was the observer (Table l). For the present purposes the most pertinent of these validity checks were the correlations that obtained between the measures of everyday memory and