Object and proper name retrieval in temporal lobe epilepsy: A study of difficulties and latencies

Highlights

• We provide a qualitative and quantitative assessment of naming difficulties in TLE.

• We focus on naming latencies and a thorough evaluation of naming errors.

• Performances were compared based on lateralization of seizures.

• Epilepsy of the hemisphere dominant for language impairs naming and reaction times.

• Both qualitative and quantitative difficulties were identified.

Summary

Purpose

Retrieving a specific name is sometimes difficult and can be even harder when pathology affects the temporal lobes. Word finding difficulties have been well documented in temporal lobe epilepsy (TLE) but analyses have mostly concentrated on the study of accuracy. Our aim here was to go beyond simple accuracy and to provide both a quantitative and a qualitative assessment of naming difficulties and latencies in patients with TLE.

Methods

Thirty-two patients with temporal lobe epilepsy (16 with epilepsy affecting the cerebral hemisphere dominant for language (D-TLE) and 16 with epilepsy affecting the cerebral hemisphere non-dominant for language (ND-TLE)) and 34 healthy matched control subjects were included in the study. The experiment involved naming 70 photographs of objects and 70 photographs of celebrities as fast as possible. Accuracy and naming reaction times were recorded. Following each trial, a questionnaire was used to determine the specific nature of each subject's difficulty in retrieving the name (e.g., no difficulty, paraphasia, tip of the tongue, feeling of knowing the name, etc). Reaction times were analysed both across subjects and across trials.

Key findings D-TLE patients showed consistent and quasi-systematic impairment compared to matched control subjects on both object and famous people naming. This impairment was characterized not only by lower accuracy but also by more qualitative errors and tip of the tongue phenomena. Furthermore, minimum reaction times were slowed down by about 70 ms for objects and 150 ms for famous people naming. In contrast, patients with ND-TLE were less impaired, and their impairment was limited to object naming.

Significance These results suggest that patients with TLE, in particular D-TLE, show a general impairment of lexical access. Furthermore, there was evidence of subtle difficulties (increased reaction times) in patients with TLE.

Introduction

Talking engages many processes, beginning with the thought of what we want to express and ending with the rapid articulation of sequences of words (Indefrey, 2011). This extraordinarily efficient and effortless ability (Alario et al., 2004) is underpinned by a widespread brain circuitry in which the temporal lobes play a central role (Hirsch et al., 2001). It is well established that people with temporal lobe epilepsy (TLE) have word finding difficulties following temporal lobe resection surgery (Drane et al., 2008). However, patients complain about word finding difficulties even before resection (Piazzini et al., 2001). Lomlomdjian et al. (2011) showed that almost one-third of patients with TLE reported frequent and severe word finding problems during spontaneous speech. These complaints are not systematically related with performance on testing (Thompson and Corcoran, 1992, Vermeulen et al., 1993, Giovagnoli et al., 1997), particularly when visual confrontation naming tasks are used (compared to auditory naming tasks, e.g. Hamberger and Tamny, 1999, Lomlomdjian et al., 2011). However, tests usually focus on accuracy and relate patients’ failures to lexical loss, i.e. anomia (Laine and Martin, 2006). A recent study attempted to overcome this limitation using a detailed procedure. Drane et al. (2013) used the Iowa Famous Faces test to show that naming deficits are consistently observed in TLE patients both before and after resection surgery. They also noted that having additional time did not help these patients and that there were qualitative differences between left and right TLE patients.

Other difficulties could explain patients’ problems. For example, even typical individuals occasionally experience difficulty retrieving a particular word, even if they know or ought to know the word. This complex cognitive phenomenon, which differs from lexical loss, is called “word retrieval failure” and has been investigated in linguistic and neuroimaging studies (Evrard, 2002, Burke and Shafto, 2004, Maril et al., 2005). There are two types of word retrieval failure. The first is the tip of the tongue phenomenon (TOT) which is defined by an incapacity to produce a specific word associated with a strong feeling that retrieval is imminent and usually combined with partial recall of some features of the word that is missing (phonological cues, word length, etc.). This phenomenon is thought to be due to the blocking of phonological representations (Brown and McNeill, 1966, Brown, 1991, Rastle and Burke, 1996, Burke and Shafto, 2004, Lindín et al., 2010). The second type of retrieval failure is the “feeling of knowing the name” phenomenon (FOKn), which is the feeling that one probably knows or has known the name but that it is too deeply buried in memory to be easily retrieved. As such, it is associated with an inability to access the lexical representation of the word. TOT and FOKn are both observed in healthy subjects (Evrard, 2002). Word retrieval failures can be solved after some time (the word one is searching for may come back to mind after a few seconds) or may remain unresolved, leading to the frustration so characteristic of these states.

Thus, a word retrieval failure is by definition characterized by abnormal latencies in name production. Studying naming latencies thus appears critical to investigating such language difficulties. However, this has seldom been carried out in previous studies of epileptic patients, a major problem since some anti-epileptic drugs are known to increase word finding difficulties. Gallegos and Tranel (2005) compared naming latencies for happy and neutral faces of celebrities in patients who had undergone left or right anterior temporal lobectomy and controls. They did not find any difference between groups, but they did not control for long latencies due to naming difficulties, such as Tip-of-the-Tongue States. Recently, Lomlomdjian et al. (2011) evaluated the time needed to correctly name pictures of the Boston Naming test in TLE patients and found no significant difference relative to controls. In contrast, mean response reaction times for famous face naming were found to be longer in both left and right pre-surgical TLE patients in Drane et al.’s (2013) study. However, the naming delays reported in these studies, as in some others (e.g., Bell et al., 2003), were longer than the expected delays reported in other studies (for a review, see Indefrey, 2011). This may be a problem since patients may use various sorts of compensatory strategies that could hide their difficulties if they are allowed unlimited time to answer. We therefore applied speed constraints in the present study to avoid this potential compensation.

Healthy subjects have more naming difficulties with proper names than common names (Evrard, 2002). In fact, the TOT phenomenon mostly affects the retrieval of proper names (Brédart, 1993). In certain cases, anomia may even be limited to proper names and spare common names (Semenza and Zettin, 1989, Semenza, 2006). Proper names are also more difficult to retrieve than other information about people (Cohen and Burke, 1994, Valentine et al., 1996). Left anterior temporal structures are known to be involved in proper name retrieval (Damasio et al., 1996, Tranel et al., 1997, Gorno-Tempini et al., 1998, Fukatsu et al., 1999, Grabowski et al., 2001, Glosser et al., 2003, Rotshtein et al., 2005, Trebuchon-Da Fonseca et al., 2009). For this reason, and because of their social importance, proper names are of special interest for people with TLE.

In this study, we aimed to obtain an objective quantification and characterization of naming difficulties in patients with TLE. To address this issue, we analysed accuracy, latencies, and the type of word retrieval difficulty (TOT, FOKn, paraphasia …) when participants named photographs of objects and famous people as fast as they could. Based on the available literature, we hypothesized that D-TLE (TLE affecting the cerebral hemisphere dominant for language) would lead to impaired naming accuracy and latency for both objects and celebrities while ND-TLE (TLE affecting the cerebral hemisphere non-dominant for language) would cause impaired accuracy and latency in naming celebrities but not objects.