The role of the neuropsychological assessment in the management of people with epilepsy has evolved considerably over the past 25 years. This paper describes some of the most common applications of a neuropsychological assessment in the diagnosis, management and treatment of people with epilepsy. It describes the factors that influence the interpretation of neuropsychological test scores in this patient group and outlines the limitations of the investigation. It gives guidelines for the optimal timing of a referral, together with timelines and indications for reassessment, and provides a checklist to help the referring clinician get the most from a neuropsychological assessment for their patients with epilepsy.
- epilepsy, surgery
Statistics from Altmetric.com
The role of the neuropsychological assessment in the management of people with epilepsy has evolved considerably over the past 25 years. Before the advent of MRI, clinicians used the patterns of cognitive strengths and weaknesses derived from a neuropsychological assessment to lateralise and localise cerebral pathology in this group. The advances in structural and functioning imaging over the past quarter of a century mean that this is no longer the primary role of a neuropsychological assessment in this population.1
However, far from hastening the demise of the neuropsychological assessment, new imaging techniques have enhanced the role of traditional tests in the clinical management of people with epilepsy.
This paper describes the ways in which a neuropsychological assessment can help in the diagnosis, management and treatment of people with epilepsy and provides some recommendations for getting the most out the investigation for this patient population.
What is a neuropsychological assessment?
A neuropsychological assessment comprises the administration of several tests that have been standardised and normed on healthy populations. Every neuropsychological test has a ceiling and a floor. These are the limits of the best and worst performances possible on the task. For example, on a ‘count to 3’ task, the ceiling of the test would be the ability to count to 3. Since well over 99% of the adult population can do this, it would be a test with a very low ceiling. Full marks on the test would not distinguish the patient whose counting limit is 4 from one who can solve complex quadratic equations in lightning speed. Most tests are therefore standardised so that the top and bottom scores are returned only by 2% of the general population. Anyone completing these standardised tests will obtain a score that will provide a measure of how well they are functioning compared with an age-matched sample of their healthy peers and/or patients with similar or other conditions. These scores tend to be normally distributed and are converted to either z scores or a percentile rank to allow comparison between tests (figure 1).
Not all neuropsychological test scores are normally distributed and some tests may be based on pass/fail criteria or have ‘cut-off’ points for the identification of an abnormal pattern of response.
Most neuropsychological assessments will generate a large number of measures of function in several cognitive domains, including general intellectual ability, memory function, expressive and receptive language skills, executive function and perceptual abilities. This creates a cognitive profile of strengths and weaknesses across multiple domains.
These cognitive profiles are examined to determine clinically significant deficits in relation to the individual’s estimated premorbid level of function in each cognitive domain (where appropriate) and also in relation to the pattern and level function seen in their healthy, age-matched peers.
When can a neuropsychological assessment help in the management of people with epilepsy?
Table 1 describes some of the most common applications of a neuropsychological assessment in the diagnosis, treatment and management of people with epilepsy.
Interpreting the results
Just as it is increasingly recognised that epilepsy is a disorder of cerebral networks2 so we have moved away from a phrenological approach of assigning specific cognitive functions to the role of specific cerebral structures.
Rather we now understand that difficulties on particular neuropsychological tasks represent dysfunction within the underlying network that subserves that function. This disruption of function can result from many factors (figure 3).
The neuropsychological assessment is unique among the standard diagnostic assessments in epilepsy. The clinical significance of the cognitive profile can only be fully interpreted in the light of the outcomes of other investigations. This is not the case with an MR scan of brain, where a cavernoma presents as a cavernoma, regardless how a patient is feeling on the day of the scan, or when they last had a seizure.
However, in a neuropsychological assessment a poor score on a test can reflect many organic and non-organic factors, some of which may be permanent (eg, underlying structural pathology), and others of which may be transient (eg, postictal disturbance, interictal subclinical electroencephalography discharges, depression, etc) or reversible (eg, medication effects) or even specific to the assessment itself (eg, poor motivation and test anxiety).
A skilled neuropsychologist must examine the evidence for each of the factors illustrated in figure 3, as well as their interactions with one another, in order to interpret the test results for each patient. This means the more information the neuropsychologist has about the patient at the time of the assessment, the better their interpretation. In most settings, neuropsychologists rely solely on the referral letter from the neurologist for this information.
Unlike a scan or an electroencephalography, a neuropsychological assessment assesses the person and not the brain. This kind of assessment has several limitations and will not yield useful results for everyone. The following limitations should be considered when considering referring someone with epilepsy for a neuropsychological assessment.
Most of the tests we use in the UK are culturally specific and have been standardised using a British population who have English as their first language. This means that the norms are valid only for this population, that is, people who have been educated within the UK and for whom English is their first language. Although there are some non-verbal tests designed to be culturally neutral, any assessment using just these tasks will be limited. While interpreters can be used to administer some tests, even qualified interpreters are not trained in the standardised administration of psychometric tests. It is possible to end up with a measure of the ‘joint IQ’ of both the patient and the translator in these situations. Even in faithful translation, the problems of the unrepresentative normative sample remain. Clinicians should therefore be cautious when interpreting the findings from a neuropsychological assessment in people who differ in important respects from the normative sample of the test.
Most neuropsychological tests are pencil and paper tasks or require the patients to answer questions put to them by the neuropsychologist (who then records their answers, with a pencil on paper). Computerised test batteries are gradually being used more widely, but technology has not transformed the neuropsychological assessment in the way that it has in other disciplines. This is because an integral part of a neuropsychological assessment is the observation of ‘how and why’ someone is failing on a task, not just whether or not they can do it, and this requires the clinician to observe closely and engage with the patient throughout the assessment. Because most of our tests rely on someone being able to see, hear and manipulate a pencil, our range of investigations is limited if someone has significant sensory or motor impairments. This does not mean that a neuropsychological assessment cannot provide useful information in these cases, but it does mean that the assessment will have to be tailored to accommodate these impairments and that it may not be possible to get a reliable measure of function in all cognitive domains.
Every neuropsychological test has a ceiling and a floor (see above). By definition, people diagnosed with learning disability will have an IQ of 70 or less and will score at or below the second percentile on tests of intellectual function. Although not always the case, it is common for function in other cognitive domains to be similarly compromised in the learning disability population. This can result in a very flat neuropsychological profile, with every score on every task falling below the second percentile. This does not mean that the individual does not have cognitive strengths and weaknesses but rather that the standardised tests are not sensitive to these patterns in this population, limiting the value of the assessment in some cases.
The patient’s motivation is by far the biggest factor that can limit the validity and reliability of neuropsychological test scores. If a patient is not willing to do their best on the tests, we will not get valid data. A neuropsychological assessment can be a tough investigation. Unless they are in the top 2% of the population, every patient will be pushed until they cannot answer the questions. As they reach the limits of their cognitive capacity, it requires increasing effort not just to give up. Patients who are anxious and depressed or who just do not see the point of the investigation will often give up before we can establish the limits of their capacity.
If a patient has undergone a full neuropsychological assessment, most neuropsychologists will advise waiting at least 9 months before attempting any reassessment. This is because practice effects can have a significant impact on performance the second (or third or fourth and so on) time around. These practice effects can mask underlying deterioration if a reassessment is carried out too soon. Exceptions to this rule are specialist serial assessments that are conducted as part of a treatment evaluation (eg, surgical follow-up): here the standardised test batteries at each stage of treatment and follow-up have been carefully designed, employing parallel test batteries to minimise practice effects. Shorter test–retest intervals can also be used in circumstances where the initial assessment was clearly compromised by factors that have since resolved, such as drug toxicity or psychological disturbance.
Routine reassessment is not advised for most people with epilepsy since overexposure results in reduced power of the tests to detect both impairment and change in function. If a patient has undergone an assessment that revealed significant memory impairment with function below the second percentile, further assessment a year down the line when they continue to complain of memory problems is unlikely to help in their management, since it is likely to show the same pattern. Unless there are other features of progressive deterioration, such a patient should be referred for memory rehabilitation rather than undergo serial assessments that are highly likely to yield the same results. Unless you think it is likely that something has changed, a reassessment is unlikely to help.
The impact of antiepileptic drugs
Given the volume of the literature on the effects of antiepileptic drugs on cognition, there are surprisingly few well-controlled studies published. As a general rule of thumb, older medications seem to have more impact on cognition than newer ones, and people taking polytherapy seem more often to be slowed up than those only taking one drug, but the impact is different for different people and there are some notable exceptions to these rules. While most antiepileptic drugs are associated with a degree of cognitive slowing, topiramate appears to have a specific impact on verbal function and there are reports of significant reductions in verbal fluency and intellectual function in some patients. A recent functional MRI study reported medication-specific effects (topiramate vs zonisamide vs levetiracetam) on the functional neuroanatomy of language and working memory networks, with topiramate and zonisamide associated with dysfunction in frontal and parietal cognitive networks and associated impaired performance on associated tasks.3 We should also consider the indirect impact of antiepileptic drugs on cognition. The mood disturbance reported by some people taking levetiracetam can also lead to poor cognitive function for some. The prescribing neurologist should carefully consider the cognitive profile of any antiepileptic drug, particularly when considering changes that could have dramatically impact on academic performance at key stages of an person’s life.
Getting the most from a neuropsychological assessment
The following checklist will help neurologists to get the most out of a neuropsychological assessment.
Think about the clinical question you want the neuropsychological assessment to address
Whether it is establishing a cognitive baseline before medication changes, capacity to consent to treatment or advice about employment options, the more specific the questions you ask in your referral, the more tailored the subsequent assessment will be. A usual rule of thumb is that the more general the referral, the more general the report you will get back. Referrals do not have to be limited to just one question—often the same results can be used in a number of ways—but it is only by asking the questions that the neuropsychologist will know what you want from the results.
Give as much information in the referral as possible
Remember that many factors contribute to a neuropsychological profile. If your neuropsychologist does not have ready access to the patient’s clinical history, include as much clinical information as possible in your referral. This information does not need to be tailored specifically to the neuropsychologist: just include copies of the most recent clinic letters. Any screening data that you have administered in the clinic (Hospital Anxiety and Depression Scale, Quality of Life in Epilepsy Scale, Addenbrookes Cognitive Examination, Montreal Cognitive Assessment) are useful but not essential, as most neuropsychological assessments will include measures in these domains.
Prepare the patient
Tell your patient why you want them to have the tests and how you think these will help. Give them some idea of what to expect. It is part of a neuropsychologist’s job to explain the nature of the tests and to establish a good rapport with the patient before they start testing them, but the neurologist can do a lot of the ground work to ensure a valid assessment by ensuring the patient is on board, at least with the rationale behind the assessment, before they arrive. If a patient arrives with a good idea of why they have been referred and how the results of the tests will be used to help manage their condition, their motivation to complete the tests to the best of their ability will be better than if they have no idea why they have been referred. Bear in mind that if you typically have difficult clinical consultations with a patient, these difficulties will be magnified under the stressful conditions of a cognitive assessment and it may not be possible to get valid results.
There are no magic bullets when it comes to cognitive dysfunction in epilepsy. Memory problems, difficulties with attention and concentration, and executive dysfunction are often an integral part of the condition because seizures and cognitive difficulties are both a manifestation of a common underlying brain pathology. To make things worse, the cognitive problems associated with the pathology are often exacerbated by the treatments aimed at reducing the seizures, be they antiepileptic medications or surgery. Clinicians often mistake this for common knowledge, but it is remarkable how many people with epilepsy report that these aspects of their condition have never been explained to them. When people with epilepsy complain of memory problems in the clinic, education about the wider manifestations of the condition should begin there and then, to ensure realistic expectations of any further investigations you may request.
Being clear about what you want from a neuropsychological assessment and how the results will affect your management of a patient before you refer will ensure that the results you get back are useful.
Prepare the patient. Preparation for the assessment begins in the neurology consultation, when you decide to refer; the better prepared the patient, the more valid the results will be.
Think carefully before you refer for a reassessment; neuropsychological assessments should be used sparingly to maintain the sensitivity of the tests.
There are no magic bullets when it comes to cognitive dysfunction in epilepsy. Ensure the patient has realistic expectations of any neuropsychological investigations that you may request.
Competing interests SB received a speaker honorarium for a presentation at an epilepsy meeting funded by UCB in November 2015.
Provenance and peer review Commissioned; externally peer reviewed. This paper was reviewed by John Paul Leach, Glasgow, UK.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.