Rehabilitation of lexical and semantic communicative impairments: An overview of available approaches

Lexical-semantic impairments are common consequences of acquired neurological damage. However, little is known about the benefits of existing treatment methods for this type of language impairment. Objective To evaluate current research into lexical-semantic interventions for adults with dementia, TBI or stroke. Methods The PubMed, PsycInfo and SCOPUS databases were searched for studies related to rehabilitation, neurological conditions, communicative and lexical-semantic skills published between 2004 and 2014. Results Twenty-eight of the 453 abstracts found were selected for the review based on the PRISMA method. Most of the studies described treatments for anomia. Semantic tasks were the most commonly used, followed by phonological and gestural strategies. Interventions were individual and involved formal tasks, although the number, frequency and duration of sessions varied between studies. Conclusion Although lexical-semantic interventions lead to improvements in language abilities, they are still poorly described in the literature, and must be further investigated in terms of their efficacy, effectiveness and long-term effects.


INTRODUCTION
L exical-semantic processing refers to language comprehension and expression at the word level. 1,2 Much of the knowledge regarding these processes has been obtained through the study of lexical-semantic impair-ments in populations with acquired neurological damage, especially left hemisphere strokes 3 and traumatic brain injury (TBI). 4 Significant research has also been conducted into the lexical-semantic abilities of individuals with neurodegenerative conditions such as dementia. 5 These studies have been particularly relevant for primary progressive aphasia (PPA), 6 whose subtypes are classified based on the nature of the language impairments present. 7 Lexical and semantic impairments may cause a variety of linguistic and cognitive alterations, the most common of which is anomia, consisting of difficulties in naming or specific word recall. 8,9 Although the left hemisphere (LH) is traditionally considered dominant for language processing, lexical and semantic deficits have also been reported in patients with right hemisphere damage (RHD). Such impairments are especially evident in semantic judgment and association tasks, 10 in the metaphorical understanding of ambiguous language, as well as in verbal fluency tasks, in which individuals with RHD tend to evoke more abstract, less prototypical and more uncommon words than control subjects. 11,12 Language is an especially important domain in neuropsychological assessment, not least due to its bidirectional relationship with other cognitive functions such as the different types of attention and memory, and executive functions. Additionally, language is also the means by which most neurocognitive instruments evaluate their target constructs. As such, language development has been the focus of several neuropsychological studies, which have found, for instance, that it may be influenced by factors such as age, education and gender even in healthy individuals. 13,14 Associations among cognitive skills, language development, cognitive stimulation and lifestyle factors must always be considered during neuropsychological evaluation and rehabilitation. Therefore, in an attempt to ensure the comprehensive assessment of language skills, several instruments have been made available for the evaluation of phonetic-phonological, syntactic-semantic and pragmatic-discursive features of spoken language, especially in patients with acquired brain injury. 15 Given the high prevalence of anomia as a symptom of neurological disorders, several interventions have also been developed to assist with the compensation or attenuation of naming difficulties based on different theoretical models. 16 Common examples of such interventions include semantic feature analysis (SFA); [17][18][19] and semantic treatments, 20,21 although interventions have also been developed based on phonological 22 and gestural approaches, 23 or semantic priming. 9 More recently, strategies involving cognitive-linguistic and communicative therapy approaches 24 and sentence generation 25 have also been described. Interestingly, although several techniques have been developed for the rehabilitation of patients with classic aphasia following LHD, 19 few em-pirical studies, reviews or meta-analyses have evaluated interventions for communicative impairments, and only general guidelines are available on the topic. 26 Literature reviews have proved to be a useful tool for the evaluation of the effects of different interventions on lexical-semantic processing in conditions such as anomic aphasia 16 or semantic variant PPA. 21 Prosodic, discursive and pragmatic impairments, as well as the treatment needs and intervention guidelines for adults with RH lesions have also been reviewed by Ferré et al. 10 Therefore, in light of the important contributions made to the literature by other reviews in the past, the aim of the present study was to use this method to describe and evaluate current research on lexical-semantic interventions, focusing on the objectives, methods and results of the studies performed on the topic. This study also entailed a careful and detailed analysis of the methods used by different lexical-semantic interventions, such as the number, frequency and duration of rehabilitation sessions, the number of therapists involved, the use of individual versus group interventions and formal or ecological tasks, as well as the minimal performance criteria applied. This information can help guide future practice and research into lexical-semantic rehabilitation. Our aim can be translated into the following research question: What methodological variables (sample, assessments and interventions), linguistic features and outcomes are evaluated in existing lexical-semantic rehabilitation programs for patients with dementia, TBI or stroke? Literature review. The present review was based on the PRISMA guidelines, 27 and performed in May 2014. The PubMed, Psycinfo and Scopus databases were searched for articles regarding lexical and semantic rehabilitation published in the past 10 years. Since interventions for patients with anomia have a long history and are among the most frequently discussed in the literature, we decided to focus on more recent studies in the area. Since most of the rehabilitation programs developed over the past ten years have been based on theoretical models and on assessment methods established in the 1990s and 2000s, we focused on studies published within this time period. 28 Different sets of keywords were used to retrieve articles related to each of the four main constructs evaluated in the present review (rehabilitation, acquired neurological damage, communicative skills and lexical-semantic abilities). The following keywords were used to retrieve articles regarding rehabilitation interventions: "rehabilitation", "readaptation", "reeducation", "training", "intervention", "treatment", "therapy", "functional recovery" and "remediation". Articles relating to neurological conditions were identified using the following terms: "stroke", "cerebrovascular disease", "cerebrovascular accident", "right hemisphere damage", "left hemisphere damage", "lesion studies", "brain injury", "brain damage", "traumatic brain injury", "closed head injury" and "dementia". Investigations into communicative skills were retrieved using the keywords "communication", "linguistic", "language", "communication", "communicative" and "aphasia". Lastly, studies of lexical-semantic processing were retrieved using the terms "lexical", "lexicon", "semantics", "verbal fluency", "lexical-semantic", "word level", "category fluency", "categorical fluency", "letter fluency", "phonemic fluency" and "semantic fluency".
The abstracts retrieved were screened based on the following inclusion criteria: 1 empirical study, 2 involvement of at least one adult with an acquired neurological condition, 3 description of lexical-semantic rehabilitation procedures, 4 presence of pre-and post-intervention assessments, and 5 publication in English, French, Spanish or Portuguese. All abstracts were examined by two independent judges, and discrepancies were settled by a third reviewer. This process resulted in the exclusion of 419 abstracts (n=334 focused on language assessment only and n=30 on general cognitive rehabilitation, n=23 were literature reviews, n=10 were duplicates, n=10 focused on medication effects, n=5 described attention and memory rehabilitation strategies, n=3 dealt with motor interventions only, n=2 evaluated the effects of music therapy, n=1 assessed the effects of psychotherapy, and =1 described occupational therapy interventions).
As can be seen in Figure 1, after exclusion criteria were applied, only 38 of the initially retrieved abstracts remained. These abstracts were reevaluated by the two judges, who disagreed on the inclusion of three articles. The third reviewer decided on the inclusion of one article, while the other two were excluded. The remaining 36 articles were read in full, resulting in the further exclusion of four papers which consisted of meta-analyses and literature reviews, as well as another four which dealt exclusively with language assessment. Consequently, a total of 28 articles were included the present review.
The objectives, sample, theoretical basis of the intervention, language assessment instruments and results described in each of the studies included are described in Table 1. Further details regarding the interventions themselves are shown in Table 2, which describes the number, frequency and duration of the rehabilitation sessions conducted, as well as the number of therapists involved in the treatment, the use of group versus individual interventions and formal versus ecological tasks,  as well as the establishment of minimal performance criteria.
Both participants improved their ability to name target nouns. Gains were maintained at follow-up, and retrieval strategy was generalized to untrained items. Functional benefits were reported.
Both methods led to similar improvements in naming ability.
To evaluate the efficacy of cognitive-linguistic and communicative treatment in aphasia after stroke.
75 patients with aphasia following left-hemisphere strokes.
Cognitive-linguistic and communicative treatment.
Both treatments led to significant improvements in communicative skills. To investigate the effects of phonological treatment for anomia.
10 patients with aphasia following left-hemisphere strokes.
Positive treatment effects were observed in confrontation naming, phonologic production and nonword repetition. Generalization to discourse production was observed. Effects remained during follow-up.
To assess the effects of semantic treatment in progressive and stroke-induced aphasia.
2 patients with progressive aphasia and 1 patient with stroke-induced aphasia. Generative naming for selected semantic categories.
To assess the efficacy of a treatment combining semantic feature analysis and semantic priming.
3 patients with anomia following left-hemisphere strokes.
SFA and semantic priming.
Treatment gains and generalization were observed in two participants. The patient with severe semantic impairment did not benefit from treatment. Naming treatment with phonological and semantic cues.
Treatment delayed the progression of language impairments.

(45)Beeson & Egnor (2005)
To assess the effects of a treatment involving both spoken and written naming in patients with lexical-semantic impairments. 2 patients with stroke-induced aphasia.
Copy and Recall Treatment (CART) and spoken repetition of selected stimuli.
Treatments involving combined written and spoken naming led to significant improvements in patients with residual phonological abilities. Semantic and phonological treatment.
Semantic and phonological treatments led to similar improvements in communicative skills.
PPA: primary progressive aphasia; SFA: semantic feature analysis; PCA: phonological components analysis.  Not specified gies such as Errorless 3,5 and Errorful Naming Treatments, 5 procedural semantic categorization tests, 34 intention and pantomime gestures 35 and Copy and Recall Treatment, which combines both spoken and written naming. 44 Only one study involved the use of communicative interventions, which aim to optimize linguistic exchanges using compensatory strategies and residual linguistic skills. 24 Some studies also performed comparisons between two or more rehabilitation strategies, in an attempt to identify which would be most suitable for the population investigated. Some of the comparisons made by the studies reviewed included face-to-face versus telerehabilitation naming treatment, 32 lexical-semantic stimulation versus unstructured cognitive stimulation, 36 and errorless naming treatment versus gestural facilitation training. 3 In addition to these studies, a further four investigations involved comparisons between multiple rehabilitation approaches. 5,20,24,37 In all of the studies reviewed, treatment stimuli were selected based on a baseline assessment. Additionally, 21.42% of studies 9,29,33,38,46 evaluated participant performance during the therapeutic process, 81.48% 3,6,9,[22][23][24][30][31][32][33][34][36][37][38][39][40]42,[44][45][46][47] reassessed patients immediately after the training program was completed and 50% 3,5,6,21,[31][32][33][36][37][38][39]41,43,44,47 performed follow-up assessments. Follow-up periods ranged from five days to six months after the end of the intervention. Linguistic competencies were assessed using formal tasks in all of the studies reviewed. Only 25% of the studies evaluated other cognitive components in addition to language, such as attention, memory and executive functioning.
Overall, post-treatment assessments indicated that the interventions led to significant improvements in linguistic performance. As can be seen in Table 2, the interventions involved between five and 96 sessions, performed one to seven days a week. A total of 42.85% of studies involved two weekly sessions. 3,5,6,9,20,23,36,37,41,43,45,46 The duration of treatment sessions ranged from 30 to 180 minutes, and in 25% of studies, treatment was performed by a single therapist, while 8% of treatments involved more than one speech therapist. 6,22 The remaining studies did not specify the number of therapists involved in the interventions described. Only one study involved group therapy, 3 with all remaining studies involving individual interventions and formal rehabilitation methods. Nine of the studies evaluated set mastery criteria for patient accuracy. 3,9,23,24,29,33,35,41,45.

DISCUSSION
The aim of the present paper was to review the existing empirical research into lexical-semantic interventions for adult patients with dementia, TBI and stroke. More specifically, this review aimed to answer the following question: What methodological variables (sample, assessments and interventions), linguistic features and outcomes are evaluated in existing lexical-semantic rehabilitation programs in dementia, TBI and stroke?
Most of the studies retrieved focused on the rehabilitation of anomia caused by strokes or neurodegenerative diseases, and involved attempts to search for the most adequate therapeutic interventions for this type of impairment. Given the wide variability in the language impairments observed across neurological conditions, most language rehabilitation research is presented in the form of case studies. Such designs can provide important data as to the effectiveness of different rehabilitation interventions, and allow for the description of interventions which are specifically tailored to the profiles of the patients evaluated. 48 Although semantic and lexical approaches are still the most commonly used for the treatment of anomia, a growing number of studies have been concerned with comparing the effectiveness of such strategies with that of other techniques, such as phonological interventions. 33 However, these comparative studies have not yet reached a consensus as to which intervention might be most appropriate for the treatment of anomia.
Other therapeutic approaches, such as the gestural facilitation of naming, were also investigated in the studies reviewed. Raymer et al., 23 for instance, found that gestural facilitation training was able to help participants recall nouns and verbs, producing beneficial results when combined with errorless naming treatment. 35. In addition to helping with word recall, the use of gestures in a therapeutic setting may encourage patients to make greater use of non-verbal strategies in daily communication, increasing their communicative competence and providing an alternative means of expression for use when word recall is impaired.
Several studies also investigated language impairments in bilingual individuals, and attempted to identify which treatments may be most beneficial for such patients. Bilingual patients with aphasia have been found to respond positively to conventional naming interventions, 6 and to benefit more from monolingual than from bilingual interventions, 39,40 especially when performed in the patient's non-native language. 40,43 Interestingly, Kurland and Falcon 39 revealed that training involving cognates, or words with similar semantic and phonological features in both of the languages spoken by the patient, may not necessarily contribute to the generalization of therapeutic benefits from one language to the other. These results suggest a possible interference effect, whereby the increased lexical access to words in one language may impair access to similar words in other languages. However, studies of bilingual patients with aphasia are still quite recent, and further investigations involving larger samples, more comprehensive assessments and longer follow-up periods must be performed to confirm these hypotheses.
One of the studies also compared the effects of conventional face-to-face therapy and computerized telerehabilitation. 32 Such studies suggest a search for intervention models which are more accessible to patients who live far from rehabilitation centers or who have locomotor disabilities. The use of such strategies may also contribute to treatment adherence and to the generalization of treatment effects.
The effectiveness of errorless and errorful learning 49 for the treatment of language impairments in patients with dementia were also compared by some of the studies reviewed. Although one study 37 found errorless learning to be a superior treatment method, other investigations found no differences between the effectiveness of the two treatments. 5 To settle these discrepancies, Class III studies must be performed to investigate the efficacy of each of these approaches. 50 Although most pre-and post-treatment assessments involved language evaluation tools only, some studies investigated the effects of language interventions on cognitive skills, especially memory, attention, and the executive functions. These studies found that lexicalsemantic interventions may lead to improvements in working memory and executive functions, 36 and that functions such as inhibitory control may contribute to the generalization of therapeutic gains. 42 These findings point to the presence of cross-domain effects of language stimulation on other cognitive abilities.
The pre-and post-treatment evaluation of cognitive abilities contributes to our comprehension of linguistic and cognitive functioning, and helps to elucidate which cognitive components may be involved in each method of lexical-semantic rehabilitation. This knowledge may help guide the planning of therapeutic interventions, and contribute to the prediction of patient prognosis and functional status following lexical-semantic therapy.
Although most of the studies analyzed produced positive results, their findings must be interpreted in light of a few limitations. The small sample size involved in the investigations limited the generalizability of their results. The variability in the tasks used to assess partic-ipant performance also prevented comparisons across studies and interventions. It is also important to note that most of the studies reviewed did not involve random assignment to treatment nor blinded pre and posttreatment assessments.
The generalization of therapeutic gains to untreated stimuli was seldom assessed in these studies, even though this variable is known to be an important indicator of the success of language interventions. In addition to evaluating performance on treated and untreated stimuli, intervention studies should also assess the effects of therapy on daily linguistic functioning. 48,51 Questionnaires to evaluate improvements in the patient's daily routine following treatment may make especially important contributions in this regard, and should be more widely implemented given that restoring patient functioning can be considered the primary aim of rehabilitation programs. There is also a need for more extensive follow-up assessments after the end of treatment, so that the long-term benefits of each intervention can be identified and compared. 50 The present review concluded that current research into lexical and semantic rehabilitation is still somewhat limited in its description of the procedures involved and the results obtained by different intervention strategies. There is a need for further studies which provide more detailed descriptions of lexical-semantic rehabilitation methods and their theoretical basis, so as to facilitate their replication by different investigators. Pre-and post-intervention assessments should also be similar across studies to allow for comparisons between investigations, and follow-up evaluations should be more carefully considered. Lastly, we suggest that the different lexical-semantic rehabilitation methods be evaluated through multicenter studies, which would allow for the participation of a larger sample.
One limitation of the present study was its exclusive focus on articles involving patients with dementia, TBI and stroke. Future reviews may also include articles discussing the effectiveness of lexical-semantic rehabilitation programs in patients with other neurological diseases such as multiple sclerosis, epilepsy, and cancer. Another limitation is the fact that the present review did not follow all PRISMA guidelines, since its scope did not include an evaluation of the quality of the studies. However, the present review did follow 20 out of the 27 items listed in the PRISMA checklist. 52 The lexical-semantic interventions described tended to make use of decontextualized stimuli. Therefore, given the facilitating effects of context on word learning and lexical access, there is a need for greater investment