Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Aug 1.
Published in final edited form as: Neuropsychol Rehabil. 2020 May 19;31(7):1145–1162. doi: 10.1080/09602011.2020.1762672

Treating limb apraxia via action semantics: a preliminary study

Harrison Stoll 1, Matthieu M de Wit 1,*, Erica L Middleton 1, Laurel J Buxbaum 1,2
PMCID: PMC7674248  NIHMSID: NIHMS1593559  PMID: 32429797

Abstract

Limb apraxia is evident in approximately 50% of patients after left hemisphere cerebral vascular accident (LCVA) and increases disability and caregiver dependence. Individuals with apraxia exhibit abnormalities in spatio-temporal aspects of gesture production and/or in knowledge of tool-related actions (action semantics). This preliminary study of three LCVA participants aimed to (i) explore the efficacy of a novel Action Network Treatment (ANT) that focused on improving the semantic association between tool actions and other types of tool knowledge, an intervention inspired by successful semantic network treatments in aphasia (e.g., Edmonds, Nadeau, & Kiran, 2009), and (ii) explore whether there are individuals with apraxia who benefit from ANT relative to a version of a comparatively well-studied existing apraxia treatment (Smania et al., 2006; Smania, Girardi, Domenicali, Lora, & Aglioti, 2000) that shapes gesture via focus on practicing the spatio-temporal aspects of gesture production (Tool Use Treatment or TUT). One participant demonstrated treatment benefits from both ANT and TUT, while another only benefited from TUT. These findings indicate that our novel semantic network strengthening approach to gesture training may be efficacious in at least some individuals with apraxia, and provide a foundation for future study of the characteristics of people with apraxia who benefit from each approach.

Keywords: Limb Apraxia, Rehabilitation, Stroke, Action, Tool Use

1. Introduction

Limb apraxia (hereafter ‘apraxia’) is a common disorder of skilled action. It is evident in approximately 50% of patients with left hemisphere cerebral vascular accident (LCVA), and also occurs in different forms in Alzheimer’s disease and corticobasal degeneration (Buxbaum et al., 2008; Zwinkels, Geusgens, van de Sande, & van Heugten, 2004). Research has identified a number of deficits associated with apraxia including, but not limited to, spatio-temporal errors in the imitation of both meaningful and meaningless movements (Buxbaum, Shapiro, & Coslett, 2014; Goldenberg, 1995, 2009) as well as deficits in real and pantomimed tool use (Gonzalez Rothi, Ochipa, & Heilman, 1991; Hermsdörfer, Li, Randerath, Roby-Brami, & Goldenberg, 2013; Watson & Buxbaum, 2015), and in semantic action knowledge (Kalénine, Buxbaum, & Coslett, 2010; Lee, Mirman, & Buxbaum, 2014; Tarhan, Watson, & Buxbaum, 2015). For example, Kalénine and colleagues demonstrated that some individuals with apraxia after LCVA make errors in identifying which of two gestures matches a semantically meaningful label (e.g., the verb “sawing”). Importantly, these deficits can occur even when individuals with apraxia exhibit unimpaired low-level motor control and strength.

In the context of a historical debate on whether the pantomime and imitation deficits associated with apraxia have real-world implications (De Renzi, 1980; Poeck, 1986), more recent research indicates that the disorder substantially impacts functional abilities. For example, individuals with apraxia after LCVA are less likely to return to work compared to those without apraxia (Saeki, Ogata, Okubo, Takahashi, & Hoshuyama, 1993), and they exhibit impairments in performance of activities of daily living and increased dependency on caregivers (Donkervoort, Dekker, & Deelman, 2006; Hanna-Pladdy, Heilman, & Foundas, 2003; Sundet, Finset, & Reinvang, 1988; van Heugten et al., 2000). Finally, research has shown that deficits of people with apraxia in producing symbolic gestures (Borod, Fitzpatrick, Helm-Estabrooks, & Goodglass, 1989) may significantly impact their ability to communicate, especially since apraxia is often co-present with aphasia (Weiss et al., 2014).

1.1. Current Treatment of Apraxia

Possibly due to the perception that the disorder has few real-world implications, little research has been devoted to the treatment of apraxia as compared to other common syndromes such as aphasia or spatial neglect resulting from CVA (Buxbaum et al., 2008; Dovern, Fink, & Weiss, 2012; Vanbellingen & Bohlhalter, 2011). A recent review paper (Buxbaum et al., 2008) noted that of the few treatments that have been studied, an apraxia rehabilitative training program (ARTP) developed by Smania and colleagues (Smania et al., 2000; 2006) appeared particularly promising (Buxbaum et al., 2008; Cantagallo, Maini, & Rumiati, 2012; Dovern et al., 2012). Over the course of several weeks, participants enrolled in the ARTP progressed through a series of tool use exercises with diminishing cues about how to perform the movements. More specifically, participants started by using real tools, were then shown pictures illustrating the gesture belonging to the tool and asked to pantomime, and finally pantomimed to tool pictures. The 6 participants enrolled showed a benefit in the ability to use real tools relative to control participants, who received standard aphasia treatment1.

In a second study, Smania et al. (2006) replicated this finding with a larger sample (N=33) and also demonstrated that the ARTP produced long-lasting performance gains. Specifically, at an initial post-treatment test, participants treated with the ARTP improved on tasks measuring praxis function, including real object use, symbolic gestures, and gesture recognition, and showed enhanced scores on the Activities of Daily Living (ADL) questionnaire. Furthermore, the benefits of the treatment were found to be persistent in that there was no decrease in ADL questionnaire scores or the tasks measuring praxis function at two months post-treatment. In both studies, Smania and colleagues found that participants treated with the ARTP showed generalization in the form of improved performance on a novel gesture task and in ADL ratings. This result is unique to the ARTP and contrasts with observations of no generalization in other apraxia treatment studies (Goldenberg & Hagmann, 1998; Goldenberg, Daumüller, & Hagmann, 2001).

1.2. A semantic approach to apraxia treatment

The ARTP focuses on improvement of spatio-temporal deficits in apraxia via motor training and practice. The spatio-temporal component of the praxis system is concerned with the planning of movement in terms of extent, direction and timing (i.e., the kinematic aspects of movement) (Buxbaum et al., 2014). In addition to the spatio-temporal component, many theoretical models of praxis include an action knowledge component (e.g., Cubelli et al., 2000; Gonzalez Rothi et al., 1991; Buxbaum, 2001; Buxbaum & Kalenine, 2010). Action knowledge includes representations of tool manipulation (the appearance of how a tool is held and moved), which may be specified in a visual format (Tarhan et al., 2015). This “core” of the gesture renders it recognizable whether large or small, rapid or slow—that is, despite the precise details of its kinematic production. Importantly, action knowledge deficits may occur in the context of preservation of other types of knowledge about tools, including their functions (Buxbaum & Saffran, 2002).

Deficits in the spatio-temporal and action knowledge components may be identified with distinct behavioral tasks. Specifically, individuals with disproportionate spatio-temporal production deficits have relatively poor performance on imitation of meaningless actions (e.g., Buxbaum et al., 2014), whereas participants with disproportionate deficits in action knowledge perform relatively poorly on gesture recognition tasks (e.g., Tarhan, et al., 2015). These deficits are neuroanatomically dissociable: spatio-temporal production deficits are associated with inferior parietal lesions, whereas action knowledge deficits are associated with posterior temporal lesions (Tarhan et al., 2015). For participants with large lesions, both components may be affected. Following from this model, we reasoned that an alternative to strengthening spatio-temporal aspects of gesture production is to strengthen action knowledge (that is, the link between tool representations and tool-use knowledge), which may improve gesture performance in at least some participants.

To our knowledge, a rehabilitation approach focusing on strengthening action knowledge is novel in the domain of apraxia. The approach is inspired by several treatments focused on strengthening the link between words and semantic information that have proven fruitful in improving naming performance in individuals with aphasia. In Semantic Feature Analysis (SFA), participants are asked on repeated trials to generate words semantically related to a target word. The rationale is that activating these related words also serves to increase activation of the target word, thereby facilitating retrieval. Results of SFA have been promising (Boyle, 2004; Coelho, McHugh, & Boyle, 2000; Kiran & Thompson, 2007) with observations of generalization of treatment effects to untreated words. The verb network strengthening treatment (VNeST) method, while similar, focuses on strengthening thematic roles (e.g., ‘agent-patient’) associated with verbs (e.g., the pair chef/sugar is associated with the verb measure). Motivation for this treatment approach comes from research suggesting that there is bidirectional activation of verbs and the people, places, and things that commonly participate in a verb’s thematic role structure (Ferretti, McRae, & Hatherell, 2001; McRae, Hare, Elman, & Ferretti, 2005). VNeST has shown benefits for picture naming, noun and verb retrieval, and picture description tasks (Edmonds & Babb, 2011; Edmonds et al., 2009).

1.3. Present study

Our primary motivation was to perform a preliminary “proof of concept” investigation of whether a semantics strengthening approach similar to the one used in SFA and VNeST might improve pantomime performance in at least some individuals with apraxia. We developed a novel treatment called the Action Network Treatment (ANT), which targets the action knowledge deficit observed in many individuals with apraxia by strengthening the associations between tools, their related actions, and other information in a putative semantic network. Participants are asked to generate the people, locations, and items commonly associated with a tool. Once provided, the responses are then repeated back to the participant while he or she performs an action (e.g., show me how a chef might use a whisk).

A second motivation for the current study was to provide preliminary evidence regarding whether individual participants might benefit from ANT as compared to a treatment focused on production of the spatio-temporal components of gesture with different types of environtmental cueing and support. We compared each participants’ improvement following ANT to a modified version of Smania et al.’s (2000; 2006) ARTP, which we called the Tool Use Treatment (TUT). The TUT required participants to engage with tools in three phases: (i) demonstration of real tool use, (ii) imitation of tool use videos (different from the original studies, in which pictures and not videos were used), (iii) pantomime to photographs. To assess generalization of both the ANT and TUT, we included sets of tools shown only during the pre- and post-treatment tests of each treatment. To increase measurement sensitivity, in both treatments we used tools selected based on their difficulty for individual participants.

2. Method

2.1. Participants

Three chronic (> 6 months post stroke onset) left hemisphere stroke participants were recruited for this study from the Neuro-Cognitive Research Registry at Moss Rehabilitation Research Institute (MRRI) (Schwartz, Brecher, Whyte, & Klein, 2005). All participants were several years post-stroke and right hand dominant prior to their stroke. To ensure that participants understood the instructions for the study, we recruited participants with no worse than moderate language comprehension deficits, defined as a score of 4 or higher on the comprehension subtest of the Western Aphasia Battery (Kertesz, 1982). Recruitment criteria were also based on apraxia severity as indexed by participant performance on a Gesture to Sight of Tools task such that their performance was at least two standard deviations below that of controls (cut-off score was 89%). Finally, participants were also selected based on their ability to commit the time necessary for completion of the study.

Table 1 reports demographic information, lesion location, Western Aphasia Battery comprehension scores, and Gesture to Sight of Tools task performance as well as performance on a Meaningless Imitation task and a Gesture Recognition task. (See Appendix B–D for task descriptions and Table 1 below for scoring information). Participant lesion drawings are provided in Appendix G. Although neither performance on gesture recognition, imitation of meaningful gestures, or lesion location informed our study inclusion criteria, we include this information in order to provide a more complete profile of the participants.

Table 1.

Participant Demographics and Neuropsychological Characteristics

Participant Age Educ. (Yrs) Mnths Post Stroke WAB Comp. Gender GtS Acc.(%) Mless Imitation Acc. (%) GR Semantic Acc. (%) GR SpatialAcc. (%) Lesion Location (Brodmann’s Areas) TLV (cubic mm)
P1 65 15 105 9 M 79.1 82.5 90 77 2, 3, 22, 34, 39, 40, 41, 42, 43, 44, 47 179606
P2 73 14 73 4.65 M 76.97 65 95 85 21, 22, 39, 40, 41, 42 64793
P3 61 15 91 9.5 F 82.97 55 92 65 1, 2, 3, 4, 6, 38, 40, 41, 42, 43, 44, 45, 46, 47 202614
Open in a new tab

Notes: WAB Comp. = Western Aphasia Battery Comprehension, (Kertesz, 1982); GtS = Gesture to Sight of Tools (Buxbaum et al., 2000); Mless Imitation = Meaningless Imitation (Buxbaum et al., 2001); GR = Gesture Recognition (Buxbaum et al., 2005); TLV = Total Lesion Volume; Listed are Brodmann Areas with > 10% damage.

2.2. Treatment item selection

Stimuli used in the treatment study were chosen from a pool of 138 tools selected from the Bank of Standardized Stimuli (Brodeur, Dionne-Dostie, Montreuil, & Lepage, 2010) and Google Images. Using the online labor market Mechanical Turk, we obtained information on five variables that have been previously shown to influence gesture performance accuracy (Watson & Buxbaum, 2018): visual complexity (n = 39), number of associated use actions (n = 37), number of moving parts (n = 26), object familiarity (n = 28), and hand posture similarity for grasping versus using (n = 28).

Items to be included in the treatment study for each participant were selected based on performance in an item selection phase. Participants were shown a 600 × 600-pixel color photograph of one of the tools from the normed set and instructed to “show me how you would use the object as if you were actually holding it in your hand”. All pantomimes were performed with the less-impaired left hand. This procedure was repeated in three separate sessions performed on different days, after which videotaped performance was coded by an independent, blinded coder. The main measure of accuracy for item selection and the pre- and post-treatment tests was a binary dependent variable termed hand posture, a measure that captures the accuracy of the shape and movement of the hand (see Appendix A). We focused primarily on hand posture (rather than arm posture, amplitude, or timing) as previous research indicates that the hand component is most sensitive to deficits in apraxia (Watson & Buxbaum, 2015).

For each participant, the 50 tools with the lowest hand posture accuracy across all three item selection sessions were selected for study inclusion. The 50 tools per participant were divided into sets corresponding to TUT treated items (n = 15), TUT untreated items (n = 10), ANT treated items (n = 15), and ANT untreated items (n = 10) while matching as closely as possible for hand posture accuracy and five normed variables (see Appendix E).

2.3. Treatments

2.3.1. Tool Use Treatment (TUT)

Based on ARTP of Smania et al. (2000), the TUT consists of three phases of progressively diminishing cues: 1) use real tools, 2) pantomime to sight of videos, 3) pantomime to sight of pictures. During each phase, participants were told what the tool was and were asked to demonstrate its use. If the experimenter deemed the attempt incorrect, they intervened with manual guidance. In the first phase, participants were instructed to pick up the tool and demonstrate its use with the phrase “Show me how you would use this”. In the second phase, participants watched a video of an actor using a tool and were prompted to “Show me how you use this if you were actually holding it and use the video to help you” while the video continued. During the final phase participants were shown a 600 × 600-pixel image of the tool and prompted to “Show me how you would use this”. Each phase of the TUT consisted of three sessions, for a total of nine sessions, with participants moving onto the next phase regardless of their performance in the previous phase. Within a session, each tool was presented three separate times. During each presentation, regardless of whether manual guidance was needed, participants were asked to demonstrate a tool’s use twice. Thus, in a single session, participants demonstrated each tool’s use six times.

2.3.2. Action Network Treatment (ANT)

Participants were first shown a 600 × 600-pixel photograph of a tool and asked a question regarding the tool. For P1 and P2 there were a total of four questions, two of which were repeated within a session for a total of 6 trials per tool presentation: “Who might use this?”, “Where might someone use this?”, “Why might someone use this?”, and “What might someone use this with?”. For P3 we added two additional questions (How and When) to ensure that each question was unique for each presentation of a specific tool. Participants provided a response (or were provided with one if unable to respond appropriately). They were then prompted to pantomime the tool’s use with a request that incorporated the response into the prompt. For example, such a sequence might be enacted as follows: “Who might use this tea kettle?” “My wife.” “Correct, show me how your wife would use this tea kettle.” As with the TUT, if the participant’s attempt was incorrect, manual guidance was provided. Participants were also presented with each tool three separate times and asked to demonstrate the tool’s use twice for each presentation. For all participants, the order of the questions was counterbalanced across sessions.

2.4. Sessions

All participants started the rehabilitation portion of the study approximately four weeks after the item selection phase. Throughout the duration of the study participants attended two one-hour sessions per week. At the beginning of the first treatment module, participants pantomimed to the sight of treated and untreated tools (25 total tools) assigned to that module to measure pre-treatment test performance. The procedure for the pre-treatment test was the same as described in section 2.2. Participants then completed the first treatment with its assigned treated tools. After nine sessions of treatment, participants completed a post-treatment test with all 25 (treated and untreated) tools assigned to that module in random order. The procedure for the post-treatment test was the same as that described for the item selection phase. Starting with the next session, participants then began the second treatment module with its set of treated and untreated tools, including a pre- and post-treatment test as described above. Order of treatment was counterbalanced across participants, with P1 and P3 receiving the ANT first and P2 the TUT.

To increase the number of trials for comparing pre- and post-treatment test performance, P3 participated in three pre- and post-treatment test sessions for each treatment (see section 2.5 Data Analysis for details).

2.5. Data Analysis

Equivalence of pre-treatment scores across the two treatments per participant was assessed via t-tests performed separately for treated and untreated items.

The analyses focused on assessing change in performance from pre- to post-treatment tests on treated and untreated items, with separate models applied to each participant’s dataset using mixed level modeling (MLM). MLM can accommodate missing and unbalanced data, affords increased power with small samples relative to ANOVA and linear regression (Goedert, Boston, & Barrett, 2013), and controls for item level variability from pre- to post-treatment test performance. Many other treatment studies have used MLM (e.g., Dignam et al., 2016; Middleton, Schwartz, Rawson, & Garvey, 2015), including treatment studies with single case designs (Malboeuf-Hurtubise, Lacourse, Herba, Taylor, & Amor, 2017; Smith, Eichler, Norman, & Smith, 2015; and see Moeyaert, Ferron, Beretvas, & Van den Noortgate, 2014; Shadish, Kyse, & Rindskopf, 2013 for reviews of single case design and MLM).

Hand posture accuracy (correct/incorrect) for each trial at the pre- and post-treatment tests were modeled using the lme4 package in R statistical software, version 3.5 (Bates, Mächler, Bolker, & Walker, 2014). The effect of a treatment on the treated or untreated items was measured by inspecting the coefficient for a two-level factor of time of test (Pre-treatment test versus Post-treatment test). For each participant, the effect of time of test for the treated and untreated items was modeled separately to quantify effects of the treatments on item-specific improvements and on generalization, respectively.

Given that P1 and P2 performed only single pre- and post-treatment tests, there was insufficient power to calculate an interaction between time of test and treatment module (McNeish & Stapleton, 2016; Shadish et al., 2013); thus, for P1 and P2, we only assessed an effect of time of test within each module. For P3, multiple pre- and post-treatment tests afforded sufficient power to assess a potential time of test by treatment interaction.

As the same items were presented for each individual at pre- and post-treatment test, in all models items were treated as a random effect except when the model failed to converge, indicative of overfitting. In those cases, we report the results from ordinary logistic regression.2

3. Results

As shown in Appendix F, there was no difference in hand posture performance at pre-treatment test for ANT versus TUT treatments for either treated or untreated items.

Figure 1 shows mean hand posture accuracy for each participant as a function of type of item (treated and untreated), time of test, and treatment (ANT and TUT). Appendix H, I, and J report coefficients and associated statistics for terms in the models for P1, P2, and P3 respectively. Note, coefficients in each model reflect an estimated change in accuracy (in log odds) from the reference level for each fixed effect. Positive coefficients indicate an increase in accuracy from the reference level to the other level of the fixed effect, and negative coefficients indicate a decrease in accuracy from the reference level to the other level of the fixed effect. For example, in Model 1 in Appendix H, the coefficient of 2.51 for the fixed effect of Time of Test reflects an increase in accuracy from Pre-treatment test (reference level) to Post-treatment test (the other level). All results reported in this section pertain only to analyses involving hand posture accuracy.

Figure 1:

Figure 1:

Open in a new tab

3.1. P1 Results

For P1, treated items following TUT showed pre- to post-treatment test improvement (53% correct vs. 93% correct), as indicated by a significant effect of time of test (Coef. = 2.51, SE=1.16, p=.03). In contrast, there was no effect of time of test for treated items for the ANT (60% correct vs. 67% correct; p = .59; Appendix H). There were no treatment effects for untreated items in either the ANT or TUT.

3.2. P2 Results

Although P2 showed numerical pre- to post-treatment test improvement for treated items with both the TUT (87% vs. 100% correct) and ANT (60% vs. 87% correct), the effects were not significant; nor were there significant improvements for untreated items (all p’s > .11; Appendix I). Inspection of P2’s performance (Figure 1) indicates that lack of treatment effects for TUT may have been influenced by high pre-treatment test scores.

3.3. P3 Results

For P3, treated items showed pre- to post-treatment test improvement following both TUT (55% vs 75% correct) and ANT (40% vs 75% correct) as indicated by a significant effect of time of test (TUT: Coef.=2.11, SE=0.82, p < .01; ANT: Coef.=2.37, SE=0.67, p < .01; Appendix J). For the untreated items, there was no effect of time of test for the TUT (p=.29). There was a significant effect of time of test for the untreated items in the ANT, but the sign of the coefficient indicates declining performance from pre- to post-treatment test (p = .02). There was no interaction between treatment and time of test for either item type (all p’s > .27).

4. Discussion

This study had two major aims: (i) to explore the efficacy of a new treatment called Action Network Treatment (ANT) that aims to treat apraxia by improving the association between tool actions, their associated tools, and other types of semantic tool knowledge, and (ii) to explore whether there are participants who appear to benefit from the semantically-oriented ANT relative to a version of a relatively well-studied existing apraxia treatment (Smania et al., 2000; 2006) focusing on shaping kinematic features of the gesture (the Tool Use Treatment or TUT).

Each of three participants showed different responses to the treatments as measured by their hand posture accuracy change from pre- to post-treatment test. For P1, hand posture accuracy for treated items improved following TUT but not ANT. For P3, hand posture accuracy for treated items improved following both treatments. P2 showed no improvement in hand posture accuracy following either treatment for any items. One reason for this could be P2’s unexpectedly high pre-treatment hand posture scores (Figure 1), which may have reduced experimental sensitivity for detecting improvement (i.e., so called “ceiling effects”). Likely due to lower pre-treatment performance (Appendix K), arm posture appeared to be a more sensitive measure of treatment effects for P2 as seen in identifiable accuracy improvements from pre- to post-treatment test for treated items following TUT and untreated items following ANT (see Appendix M). Thus, the lack of a treatment effect relating to hand posture for P2 may have been due to vagaries of item selection rather than lack of efficacy of the treatments examined.

The results for P3 suggest that the ANT holds promise for treating apraxia in some individuals, at least for treated items. Returning to the aphasia literature that inspired the ANT, we speculate that the core aspects of SFA and VNeST that we attempted to preserve in the ANT could be related to its success. Specifically, all of these treatments require participants to generate semantic features about items, which has been argued to facilitate deep semantic processing (Raymer et al., 2008). In SFA, participants are required to name a target picture and then to produce words that are semantically related to the target; these responses are then placed around the target picture in a chart (Boyle, 2004, 2010; Tompkins, Fassbinder, Scharp, & Meigh, 2008). Features include categories, use, action, location, and associations. Similarly, in VNeST, participants are shown a verb (e.g., measure) and asked to generate who (e.g., chef, carpenter) and what (e.g., scissors, knife) might be associated with the verb (Edmonds & Babb, 2011; Edmonds et al., 2009). Both treatments have proven beneficial to both treated and untreated items. In the ANT, we attempted to replicate this design by asking participants to generate semantic associations with tools. As discussed in the introduction, we focused on facilitating deep semantic processing by requiring participants to respond to a number of semantic questions about tools.

A second aspect of the ANT that may have been beneficial was its incorporation of repeated practice in retrieving tool actions in response to a cue. Outside of the domain of aphasia, benefits of retrieval practice have been demonstrated in conceptual and visual spatial learning (Kang, McDaniel, & Pashler, 2011; Karpicke & Blunt, 2011). Furthermore, research has shown that while both apraxia and aphasia are associated with semantic deficits, people with apraxia are disproportionately impaired in semantic knowledge of the actions associated with tools as compared with other types of semantic tool knowledge (Buxbaum & Saffran, 2002; Myung et al., 2010). Thus, our approach focused specifically on strengthening action semantics both by virtue of associating action semantics with other aspects of semantic knowledge as well as by incorporating retrieval practice. It will be important in future research to delineate the active ingredients of semantic-based treatments for apraxia, such as the ANT, to improve treatment efficacy and efficiency, and promote an understanding of how the treatment exerts change in specific processing deficits relating to apraxia.

The analyses on hand posture showed no evidence of generalization in the form of improvement of untreated items from pre- to post-treatment. In fact, for one participant (P3), performance on untreated items decreased from pre- to post-treatment (ANT module). In only one instance did we observe improvement on untreated items from pre- to post-treatment--following ANT for P2 in the analysis of arm posture (Appendix M). However, we are reluctant to interpret this finding as evidence of generalization because arm posture scores for treated items did not improve following ANT for P2. That is, one should expect patterns of true generalization to track patterns of direct treatment effects (i.e., on treated items).

The lack of generalization found in the current study contrasts with the results that Smania et al. (2000; 2006) reported for the ARTP. However, our findings were in line with other apraxia treatment studies that found improvement only on treated items (Goldenberg & Hagmann, 1998; Goldenberg, et al., 2001). Goldenberg et al. concluded that their training only helped participants understand a specific set of instructions for a given task and was not effective in helping participants transfer their new knowledge to novel situations. One of the key features of the ARTP is the focus on a wide variety of gestures, including both symbolic and meaningless gestures. It is possible that the engagement with gestures in a variety of circumstances allowed for better generalization to untreated gestures.

The lack of generalization in the current study also departs from the aphasia treatment studies that informed the ANT. One reason for the disparity may pertain to the nature of the untreated items. In VNeST, Edmonds and colleagues selected untreated items that were semantically related to the treated items. Similarly, in a post-hoc analysis, Boyle (2004) found that effects on untreated items were present only for semantic categories that had been accessed during treatment. In contrast, we chose untreated ANT items based on item selection performance, without regard to semantic category. Thus, any strengthening of the semantic networks of the treated items might not be expected to generalize to untreated items. A direction for future research may be to manipulate or control the semantic relatedness of treated and untreated item sets to provide a stronger test of the generalization potential of semantic-based treatments of apraxia.

Another possible future direction for advancing an understanding of the clinical potential of semantics-based (e.g., ANT) and spatio-temporal focused (e.g., TUT) treatments of apraxia may involve examination of how the specific processing deficits underlying a participant’s apraxia relates to treatment efficacy. Though our participants were well characterized, the design and small sample did not enable us to draw conclusions regarding the relationship of underlying deficits to treatment efficacy. This would be afforded by a larger study involving a case series approach, wherein the relationship between participant variability in neuropsychological and neuroanatomical profile and treatment efficacy can be formally assessed.

In summary, we have demonstrated that a novel treatment focused on strengthening the semantic association between use actions, tools, and other types of semantic information may be of benefit in at least some individuals with apraxia from stroke. Moreover, we have provided converging evidence demonstrating efficacy of an apraxia treatment that involves shaping kinematic features of gesture production with diminishing environmental cues. Given the dearth of apraxia rehabilitation research and the real-world consequences of the disorder, the findings of the present study are encouraging with respect to future investigations of apraxia treatment.

Supplementary Material

Supplemental Material

Acknowledgements.

We would like to thank Cortney Howard, Leyla Tarhan, and Louisa Smith for coding participants’ gestures; and H. Branch Coslett for help with lesion segmentation. This research was supported by NIH grant R01 NS099061 to L.J. Buxbaum and an Albert Einstein Society Award to M.M. de Wit.

Footnotes

Author Information: The authors declare no competing financial interests.

Data Availability. In compliance with the guidelines of the Institutional Review Board of Einstein Healthcare Network, all participants gave informed consent and were compensated for travel expenses and participation. The informed consents obtained did not include permission to make data publicly available; as such, the conditions of our ethical approval do not permit anonymized study data to be publicly archived. To obtain access to the data, individuals should contact the corresponding author. Requests for data are assessed and approved by the Institutional Review Board of Einstein Healthcare Network.

1

The ARTP also included a second symbolic gesture section and a third meaningless imitation section but not all participants were able to complete these sections. Hereafter, we refer to the version completed by all subjects as the ARTP.

2

Following the suggestion of a reviewer, we also performed a post-hoc analysis of arm posture scores using the same analysis approach as described for hand posture scores (section 2.5). The arm posture results are reported in Appendix L, M, and N for P1, P2, and P3 respectively. Significant treatment effects were almost altogether absent in the models of arm posture scores with the exception of two findings for P2, which we examine further in the Discussion section.

References

  1. Bates D, Mächler M, Bolker B, & Walker S (2014). Fitting Linear Mixed-Effects Models using lme4. ArXiv:1406.5823 [Stat]. Retrieved from http://arxiv.org/abs/1406.5823 [Google Scholar]
  2. Borod JC, Fitzpatrick PM, Helm-Estabrooks N, & Goodglass H (1989). The relationship between limb apraxia and the spontaneous use of communicative gesture in aphasia. Brain and Cognition, 10(1), 121–131. 10.1016/0278-2626(89)90079-1 [DOI] [PubMed] [Google Scholar]
  3. Boyle M (2004). Semantic Feature Analysis Treatment for Anomia in Two Fluent Aphasia Syndromes. American Journal of Speech-Language Pathology, 13(3), 236–249. 10.1044/1058-0360(2004/025) [DOI] [PubMed] [Google Scholar]
  4. Boyle M (2010). Semantic Feature Analysis Treatment for Aphasic Word Retrieval Impairments: What’s in a Name? Topics in Stroke Rehabilitation, 17(6), 411–422. 10.1310/tsr1706-411 [DOI] [PubMed] [Google Scholar]
  5. Brodeur MB, Dionne-Dostie E, Montreuil T, & Lepage M (2010). The Bank of Standardized Stimuli (BOSS), a New Set of 480 Normative Photos of Objects to Be Used as Visual Stimuli in Cognitive Research. PLoS ONE, 5(5), e10773. 10.1371/journal.pone.0010773 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Buxbaum LJ, Giovannetti T, & Libon D (2000). The Role of the Dynamic Body Schema in Praxis: Evidence from Primary Progressive Apraxia. Brain and Cognition, 44(2), 166–191. 10.1006/brcg.2000.1227 [DOI] [PubMed] [Google Scholar]
  7. Buxbaum LJ, Haaland KY, Hallett M, Wheaton L, Heilman KM, Rodriguez A, & Rothi LJG (2008). Treatment of Limb Apraxia: Moving Forward to Improved Action. American Journal of Physical Medicine & Rehabilitation, 87(2), 149–161. 10.1097/PHM.0b013e31815e6727 [DOI] [PubMed] [Google Scholar]
  8. Buxbaum LJ, Kyle K, Grossman M, & Coslett B (2007). Left Inferior Parietal Representations for Skilled Hand-Object Interactions: Evidence from Stroke and Corticobasal Degeneration. Cortex, 43(3), 411–423. 10.1016/S0010-9452(08)70466-0 [DOI] [PubMed] [Google Scholar]
  9. Buxbaum LJ, Kyle KM, & Menon R (2005). On beyond mirror neurons: Internal representations subserving imitation and recognition of skilled object-related actions in humans. Cognitive Brain Research, 25(1), 226–239. 10.1016/j.cogbrainres.2005.05.014 [DOI] [PubMed] [Google Scholar]
  10. Buxbaum LJ, & Saffran EM (2002). Knowledge of object manipulation and object function: Dissociations in apraxic and nonapraxic subjectsq. Brain and Language, 21. [DOI] [PubMed] [Google Scholar]
  11. Buxbaum LJ, Shapiro AD, & Coslett HB (2014). Critical brain regions for tool-related and imitative actions: A componential analysis. Brain, 137(7), 1971–1985. 10.1093/brain/awu111 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Cantagallo A, Maini M, & Rumiati RI (2012). The cognitive rehabilitation of limb apraxia in patients with stroke. Neuropsychological Rehabilitation, 22(3), 473–488. 10.1080/09602011.2012.658317 [DOI] [PubMed] [Google Scholar]
  13. Coelho CA, McHugh RE, & Boyle M (2000). Semantic feature analysis as a treatment for aphasic dysnomia: A replication. Aphasiology, 14(2), 133–142. 10.1080/026870300401513 [DOI] [Google Scholar]
  14. De Renzi E (1980). Imitating Gestures: A Quantitative Approach to Ideomotor Apraxia. Archives of Neurology, 37(1), 6. 10.1001/archneur.1980.00500500036003 [DOI] [PubMed] [Google Scholar]
  15. Dignam J, Copland D, Rawlings A, O’Brien K, Burfein P, & Rodriguez AD (2016). The relationship between novel word learning and anomia treatment success in adults with chronic aphasia. Neuropsychologia, 81, 186–197. 10.1016/j.neuropsychologia.2015.12.026 [DOI] [PubMed] [Google Scholar]
  16. Donkervoort M, Dekker J, & Deelman B (2006). The course of apraxia and ADL functioning in left hemisphere stroke patients treated in rehabilitation centres and nursing homes. Clinical Rehabilitation, 20(12), 1085–1093. 10.1177/0269215506071257 [DOI] [PubMed] [Google Scholar]
  17. Dovern A, Fink GR, & Weiss PH (2012). Diagnosis and treatment of upper limb apraxia. Journal of Neurology, 259(7), 1269–1283. 10.1007/s00415-011-6336-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Edmonds LA, & Babb M (2011). Effect of Verb Network Strengthening Treatment in Moderate-to-Severe Aphasia. American Journal of Speech-Language Pathology, 20(2), 131–145. 10.1044/1058-0360(2011/10-0036) [DOI] [PubMed] [Google Scholar]
  19. Edmonds LA, Nadeau SE, & Kiran S (2009). Effect of Verb Network Strengthening Treatment (VNeST) on lexical retrieval of content words in sentences in persons with aphasia. Aphasiology, 23(3), 402–424. 10.1080/02687030802291339 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Ferretti TR, McRae K, & Hatherell A (2001). Integrating Verbs, Situation Schemas, and Thematic Role Concepts. Journal of Memory and Language, 44(4), 516–547. 10.1006/jmla.2000.2728 [DOI] [Google Scholar]
  21. Goedert KM, Boston RC, & Barrett AM (2013). Advancing the Science of Spatial Neglect Rehabilitation: An Improved Statistical Approach with Mixed Linear Modeling. Frontiers in Human Neuroscience, 7. 10.3389/fnhum.2013.00211 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Goldenberg G (1995). Imitating gestures and manipulating a mannikin—The representation of the human body in ideomotor apraxia. Neuropsychologia, 33(1), 63–72. 10.1016/0028-3932(94)00104-W [DOI] [PubMed] [Google Scholar]
  23. Goldenberg G (2009). Apraxia and the parietal lobes. Neuropsychologia, 47(6), 1449–1459. 10.1016/j.neuropsychologia.2008.07.014 [DOI] [PubMed] [Google Scholar]
  24. Goldenberg G, Daumüller M, & Hagmann S (2001). Assessment and therapy of complex activities of daily living in apraxia. Neuropsychological rehabilitation, 11(2), 147–169. [Google Scholar]
  25. Goldenberg G, & Hagmann S (1998). Therapy of activities of daily living in patients with apraxia. Neuropsychological Rehabilitation, 8, 123–142. [Google Scholar]
  26. Gonzalez Rothi LJ, Ochipa C, & Heilman KM (1991). A Cognitive Neuropsychological Model of Limb Praxis. Cognitive Neuropsychology, 8(6), 443–458. 10.1080/02643299108253382 [DOI] [Google Scholar]
  27. Hanna-Pladdy B, Heilman KM, & Foundas AL (2003). Ecological implications of ideomotor apraxia: Evidence from physical activities of daily living. Neurology, 60(3), 487–490. 10.1212/WNL.60.3.487 [DOI] [PubMed] [Google Scholar]
  28. Hermsdörfer J, Li Y, Randerath J, Roby-Brami A, & Goldenberg G (2013). Tool use kinematics across different modes of execution. Implications for action representation and apraxia. Cortex, 49(1), 184–199. 10.1016/j.cortex.2011.10.010 [DOI] [PubMed] [Google Scholar]
  29. Kalénine S, Buxbaum LJ, & Coslett HB (2010). Critical brain regions for action recognition: Lesion symptom mapping in left hemisphere stroke. Brain, 133(11), 3269–3280. 10.1093/brain/awq210 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Kalénine S, Shapiro AD, Flumini A, Borghi AM, & Buxbaum LJ (2014). Visual context modulates potentiation of grasp types during semantic object categorization. Psychonomic bulletin & review, 21(3), 645–651. doi: 10.3758/s13423-013-0536-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Kang SHK, McDaniel MA, & Pashler H (2011). Effects of testing on learning of functions. Psychonomic Bulletin & Review, 18(5), 998–1005. 10.3758/s13423-011-0113-x [DOI] [PubMed] [Google Scholar]
  32. Karpicke JD, & Blunt JR (2011). Retrieval Practice Produces More Learning than Elaborative Studying with Concept Mapping. Science, 331(6018), 772–775. 10.1126/science.1199327 [DOI] [PubMed] [Google Scholar]
  33. Kertesz A (1982). Western aphasia battery. New York: Grune & Stratton. [Google Scholar]
  34. Kiran S, & Thompson CK (2007). The Role of Semantic Complexity in Treatment of Naming Deficits: Training Semantic Categories in Fluent Aphasia by Controlling Exemplar Typicality. 21. [DOI] [PubMed] [Google Scholar]
  35. Lee C-I, Mirman D, & Buxbaum LJ (2014). Abnormal dynamics of activation of object use information in apraxia: Evidence from eyetracking. Neuropsychologia, 59, 13–26. 10.1016/j.neuropsychologia.2014.04.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Malboeuf-Hurtubise C, Lacourse E, Herba C, Taylor G, & Amor LB (2017). Mindfulness-based Intervention in Elementary School Students With Anxiety and Depression: A Series of n-of-1 Trials on Effects and Feasibility. Journal of Evidence-Based Complementary & Alternative Medicine, 22(4), 856–869. 10.1177/2156587217726682 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Martin N, Schwartz MF, & Kohen FP (2006). Assessment of the ability to process semantic and phonological aspects of words in aphasia: A multi‐measurement approach. Aphasiology, 20(2–4), 154–166. 10.1080/02687030500472520 [DOI] [Google Scholar]
  38. McNeish DM, & Stapleton LM (2016). The Effect of Small Sample Size on Two-Level Model Estimates: A Review and Illustration. Educational Psychology Review, 28(2), 295–314. 10.1007/s10648-014-9287-x [DOI] [Google Scholar]
  39. McRae K, Hare M, Elman JL, & Ferretti T (2005). A basis for generating expectancies for verbs from nouns. Memory & Cognition, 33(7), 1174–1184. 10.3758/BF03193221 [DOI] [PubMed] [Google Scholar]
  40. Middleton EL, Schwartz MF, Rawson KA, & Garvey K (2015). Test-enhanced learning versus errorless learning in aphasia rehabilitation: Testing competing psychological principles. Journal of Experimental Psychology: Learning, Memory, and Cognition, 41(4), 1253–1261. 10.1037/xlm0000091 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Mirman D, Zhang Y, Wang Z, Coslett HB, & Schwartz MF (2015). The ins and outs of meaning: Behavioral and neuroanatomical dissociation of semantically-driven word retrieval and multimodal semantic recognition in aphasia. Neuropsychologia, 76, 208–219. 10.1016/j.neuropsychologia.2015.02.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Moeyaert M, Ferron JM, Beretvas SN, & Van den Noortgate W (2014). From a single-level analysis to a multilevel analysis of single-case experimental designs. Journal of School Psychology, 52(2), 191–211. 10.1016/j.jsp.2013.11.003 [DOI] [PubMed] [Google Scholar]
  43. Myung J, Blumstein SE, Yee E, Sedivy JC, Thompson-Schill SL, & Buxbaum LJ (2010). Impaired access to manipulation features in Apraxia: Evidence from eyetracking and semantic judgment tasks. Brain and Language, 112(2), 101–112. 10.1016/j.bandl.2009.12.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Poeck K (1986). The Clinical Examination of Motor Apraxia. In Methods in Neuropsychology (pp. 129–134). 10.1016/B978-0-08-032026-7.50011-3 [DOI] [PubMed] [Google Scholar]
  45. Raymer AM, Beeson P, Holland A, Kendall D, Maher LM, Martin N, … Gonzalez Rothi LJ (2008). Translational Research in Aphasia: From Neuroscience to Neurorehabilitation. Journal of Speech, Language, and Hearing Research, 51(1). 10.1044/1092-4388(2008/020) [DOI] [PubMed] [Google Scholar]
  46. Saeki S, Ogata H, Okubo T, Takahashi K, & Hoshuyama T (1993). Factors influencing return to work after stroke in Japan. Stroke, 24(8), 1182–1185. 10.1161/01.STR.24.8.1182 [DOI] [PubMed] [Google Scholar]
  47. Schwartz MF, Brecher AR, Whyte J, & Klein MG (2005). A patient registry for cognitive rehabilitation research: a strategy for balancing patients’ privacy rights with researchers’ need for access. Archives of Physical Medicine and Rehabilitation, 86(9), 1807e1814. 10.1016/j.apmr.2005.03.009. [DOI] [PubMed] [Google Scholar]
  48. Shadish WR, Kyse EN, & Rindskopf DM (2013). Analyzing data from single-case designs using multilevel models: New applications and some agenda items for future research. Psychological Methods, 18(3), 385–405. 10.1037/a0032964 [DOI] [PubMed] [Google Scholar]
  49. Smania N, Aglioti SM, Girardi F, Tinazzi M, Fiaschi A, Cosentino A, & Corato E (2006). Rehabilitation of limb apraxia improves daily life activities in patients with stroke. Neurology, 67(11), 2050–2052. 10.1212/01.wnl.0000247279.63483.1f [DOI] [PubMed] [Google Scholar]
  50. Smania N, Girardi F, Domenicali C, Lora E, & Aglioti S (2000). The rehabilitation of limb apraxia: A study in left-brain–damaged patients. Archives of Physical Medicine and Rehabilitation, 81(4), 379–388. 10.1053/mr.2000.6921 [DOI] [PubMed] [Google Scholar]
  51. Smith JD, Eichler WC, Norman KR, & Smith SR (2015). The Effectiveness of Collaborative/Therapeutic Assessment for Psychotherapy Consultation: A Pragmatic Replicated Single-Case Study. Journal of Personality Assessment, 97(3), 261–270. 10.1080/00223891.2014.955917 [DOI] [PubMed] [Google Scholar]
  52. Sundet K, Finset A, & Reinvang I (1988). Neuropsychological predictors in stroke rehabilitation. Journal of Clinical and Experimental Neuropsychology, 10(4), 363–379. 10.1080/01688638808408245 [DOI] [PubMed] [Google Scholar]
  53. Tarhan LY, Watson CE, & Buxbaum LJ (2015). Shared and Distinct Neuroanatomic Regions Critical for Tool-related Action Production and Recognition: Evidence from 131 Left-hemisphere Stroke Patients. Journal of Cognitive Neuroscience, 27(12), 2491–2511. 10.1162/jocn_a_00876 [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Tompkins CA, Fassbinder W, Scharp VL, & Meigh KM (2008). Activation and maintenance of peripheral semantic features of unambiguous words after right hemisphere brain damage in adults. Aphasiology, 22(2), 119–138. 10.1080/02687030601040861 [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. van Heugten CM, Dekker J, Deelman BG, van Dijk AJ, Stehmann-Saris FC, & Kinebanian A (2000). Measuring disabilities in stroke patients with apraxia: A validation study of an observational method. Neuropsychological Rehabilitation, 10(4), 401–414. 10.1080/096020100411989 [DOI] [Google Scholar]
  56. Vanbellingen T, & Bohlhalter S (2011). Apraxia in neurorehabilitation: Classification, assessment and treatment. Neurorehabilitation, (2), 91–98. 10.3233/NRE-2011-0637 [DOI] [PubMed] [Google Scholar]
  57. Watson CE, & Buxbaum LJ (2015). A distributed network critical for selecting among tool-directed actions. Cortex, 65, 65–82. 10.1016/j.cortex.2015.01.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Weiss PH, Ubben SD, Kaesberg S, Kalbe E, Kessler J, Liebig T, & Fink GR (2014). Where language meets meaningful action: A combined behavior and lesion analysis of aphasia and apraxia. Brain Structure and Function, 221(1), 563–576. 10.1007/s00429-014-0925-3 [DOI] [PubMed] [Google Scholar]
  59. Zwinkels A, Geusgens C, van de Sande P, & van Heugten C (2004). Assessment of apraxia: Inter-rater reliability of a new apraxia test, association between apraxia and other cognitive deficits and prevalence of apraxia in a rehabilitation setting. Clinical Rehabilitation, 18(7), 819–827. 10.1191/0269215504cr816oa [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental Material
NIHMS1593559-supplement-Supplemental_Material.docx (6MB, docx)

RESOURCES