Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Nov 1.
Published in final edited form as: Cortex. 2019 Jun 19;120:269–283. doi: 10.1016/j.cortex.2019.05.021

Reduced competition between tool action neighbors in left hemisphere stroke

Frank E Garcea a,b,*,1, Harrison Stoll a,1, Laurel J Buxbaum a,c
PMCID: PMC6951425  NIHMSID: NIHMS1536109  PMID: 31352237

Abstract

When pantomiming the use of tools, patients with limb apraxia after left hemisphere stroke (LCVA) produce more spatiotemporal hand action errors with tools associated with conflicting actions for use versus grasp-to-pick-up (e.g., corkscrew) than tools having a single action for both use and grasp (e.g., hammer). There are two possible accounts for this pattern of results. Reduced performance with ‘conflict’ tools may simply reflect weakened automaticity of use action activation, which is evident only when the use and grasp actions are not redundant. Alternatively, poor use performance may reflect a reduced ability of appropriate tool use actions to compete with task-inappropriate action representations. To address this issue, we developed a Stroop-like experiment in which 21 LCVA and 8 neurotypical participants performed pantomime actions in blocks containing two tools that were similar (“neighbors”) in terms of hand action or function, or were unrelated on either dimension. In a congruent condition, they pantomimed the use action associated with the visually presented tool, whereas in an incongruent condition, they pantomimed the use action for the other tool in the block. Relative to controls and other task conditions, LCVA participants showed reductions in hand action errors in incongruent relative to congruent action trials; furthermore, the degree of reduction in this incongruence effect was related to the participants’ susceptibility to grasp-on-use conflict in a separate test of pantomime to the sight of tools. Support vector regression lesion-symptom mapping analyses identified the left inferior frontal gyrus, supramarginal gyrus, and superior longitudinal fasciculus as core neuroanatomical sites associated with abnormal performance. Collectively, the results indicate that weakened activation of tool use actions in limb apraxia gives rise to reduced ability of these actions to compete for task-appropriate selection when competition arises within single tools (grasp-on-use conflict) as well as between two tools (reduced neighborhood effects).

Keywords: Limb apraxia, Tool use, Action selection, Left supramarginal gyrus, Left inferior frontal gyrus

1. Introduction

On a moment-to-moment basis our cognitive and neural systems need to integrate perceptual input with motor production processes in the service of object-directed action. In the study of language, vision, and memory, biased competition–that is, the weighting of inputs to enable selection among competing alternatives in accordance with task goals–is hypothesized to be a critical mechanism by which this process is achieved (e.g., see Desimone & Duncan, 1995; Humphreys, 1998; Kan & Thompson-Schill, 2004; Kimberg & Farah, 1993; Levelt, Roelofs, & Meyer, 1999). Skilled tool use is a defining achievement of human cognition, and poses a special challenge for selection: Many tools can be used with more than one action, and many everyday tasks require more than one tool. In that context, a range of empirical and computational studies provide evidence that competition between potential actions is biased by goals and context to facilitate task-appropriate selection (e.g., see Cisek & Kalaska, 2005; Kim & Shadlen, 1999; McKinstry, Dale, & Spivey, 2008; Pastor-Bernier & Cisek, 2011; Rounis & Humphreys, 2015; for review, see Cisek & Kalaska, 2010). Perhaps surprisingly, there is relatively little research addressing the factors that influence whether particular tool actions will compete successfully with one another or the likelihood that a given tool action will be selected.

In the domains of language and semantic memory, considerable evidence suggests that the sharing of semantic (e.g., visual or propositional) or phonological features influences competition. For example, the word “robin” may interfere with the production of the word “ostrich” by virtue of the sharing of visual and propositional feature “has wings” (e.g., see Collins & Loftus, 1975; Fieder, Wartenburger, & Abdel Rahman, 2018; Mirman & Magnuson, 2008; Rose, Aristei, Melinger, & Abdel Rahman, 2018; Vieth, McMahon, & de Zubicaray, 2014; Vigliocco, Vinson, Lewis, & Garrett, 2004), as representations with shared features are more likely to compete for selection (Damian, Vigliocco, & Levelt, 2001; Luce & Pisoni, 1998; Magnuson, Dixon, Tanenhaus, & Aslin, 2007; Vigliocco et al., 2004; for review, see; Chen & Mirman, 2012). However, recent computational modeling work suggests that when activation of a typically competitive neighbor is weak, there is a net facilitative effect on target selection; by contrast, when activation levels of a competitive neighbor are strong, there will be a net inhibitory effect on target selection (Chen & Mirman, 2012). Thus, one aim of the present investigation was to determine whether principles derived from computational modeling in the language domain extend to competition between neighboring gestures in the skilled action domain.

In the skilled action domain, eye-tracking evidence has shown that distractor tools (e.g., corkscrew) that are used with actions similar to targets (e.g., key; both used with a twisting action) cause more competition for selection than tools that are dissimilar in action (e.g., scissors), even controlling for other semantic and visual factors, and even when actions are irrelevant to the target selection task (e.g., see Lee, Middleton, Mirman, Kalenine, & Buxbaum, 2013; see also Myung, Blumstein, & Sedivy, 2006). One account of these data is that thinking about or viewing tools causes activation of actions associated with their use (e.g., see Bub & Masson, 2012), and similarity between those actions, in turn, causes the associated tools to compete with one another for selection (Buxbaum, 2017). But what determines the similarity of action features, and thus the degree to which actions will compete? Functional magnetic resonance imaging (fMRI) research in neurotypical adults provides additional resolution on this issue, as it has been suggested that actions may be organized based on visual or motor properties (e.g., environmental context or body parts used, respectively; see Tucciarelli, Wurm, Baccolo, & Lingnau, 2019; Wurm & Caramazza, 2019), and based on features less tied to the modality through which the action stimulus was presented (e.g., the abstract function or goal of the action; inferring the intention of an actor performing an action; for review, see Lingnau & Downing, 2015; Martin, 2016). However, it remains poorly understood how these additional features contribute to the organization of action similarity and whether they compete for selection.

Our laboratory (Watson & Buxbaum, 2014) previously demonstrated that tool action “neighborhoods” are likely to be determined by at least two spatiotemporal features associated with tool use, namely, the magnitude of arm movement and the posture/movement of the hand. In a follow-up study, neurotypical participants who performed a picture-word matching task achieved lower accuracy and increased response time with items that were close versus far action neighbors, as defined by magnitude of arm movement and hand action (Watson & Buxbaum, 2014). Thus, even though action features were irrelevant to the task, close action neighbors interfered with the processing of the target stimulus. Similar effects have been observed with close semantic neighbors in language and memory domains (e.g., Wei & Schnur, 2016).

1.1. Abnormal tool use competition in apraxia

Across a number of experiments, research in patients with limb apraxia–a common, highly disabling, and puzzling disorder of tool use action frequently observed after left hemisphere stroke (e.g., see Johnson-Frey, 2004; Rothi, Ochipa, & Heilman, 1991) – suggests that errors in tool-directed action are the result of specific, clinically significant abnormalities in the processes by which tool use actions are activated. For example, many apraxics produce substantially more action errors with tools for which there is a difference between the hand actions for using and grasping-to-move (e.g., corkscrew; hereafter ‘conflict’ tools) than with tools requiring the same action for moving and using (e.g., hammer; hereafter ‘non-conflict’ tools; e.g., Jax & Buxbaum, 2013; Watson & Buxbaum, 2015). For ‘conflict’ tools, the grasp-to-move action is closely aligned with prominent “affordances”, such that grasping processes can be driven by tools’ structural attributes, whereas tool use actions must be retrieved from memory. Moreover, the pattern of impaired performance with ‘conflict’ tools is associated with lesions to the left insula, inferior frontal gyrus (IFG), supramarginal gyrus (SMG), and superior longitudinal fasciculus (SLF; e.g., see Watson & Buxbaum, 2015; see also Goldenberg, Hermsdorfer, Glindemann, Rorden, & Karnath, 2007; Randerath, Goldenberg, Spijkers, Li, & Hermsdorfer, 2010). Notably, despite disrupted use action representations, apraxics perform relatively normally on tasks requiring access to knowledge of tool function (i.e., purpose; Buxbaum & Saffran, 2002).

One possibility is that reduced hand action performance with ‘conflict’ tools may simply reflect weakened use action activation, which is evident only when the use and grasp actions are not redundant. Consider, for example, that grasping a cup to use it entails the same hand action as grasping to move it, such that access to the latter information may enable an apparently correct response. However, this account does not accommodate the evidence that apraxic patients exhibit abnormalities in action competition. Specifically, eye-tracking tasks have shown that individuals with apraxia exhibit reduced competition between tools that are used with similar hand actions (e.g., see Lee, Mirman, & Buxbaum, 2014; Myung et al., 2010). Moreover, variability in the magnitude of slow and weakened competition from tool use neighbors is correlated with the degree of overt impairment in an independent gesture production task, indicating that subtle competition abnormalities reflected in eye movements bear on deficits in overt gesture production (Lee et al., 2014). Thus, one further possibility is that reduced automaticity of tool use action activation has consequences for action competition and selection in the apraxia syndrome.

1.2. Motivation for the present study

If it is indeed the case that apraxia may be associated with abnormalities in the competitive process by which task-relevant use actions are selected, then we expect to observe relatively better performance by apraxic individuals in a task in which competition from automatically-activated use actions are normally disruptive. The quintessential task with those characteristics is the Stroop task. In the classic Stroop task, a prepotent response (e.g., to read words) must be inhibited in favor of a less automatically activated, task-instructed response (e.g., to name the color of the ink in which the words are printed). Using similar logic, we predicted that LCVA participants with weak and/or slowed automatic activation of use actions with familiar tool stimuli would perform relatively well when required to inhibit the (weakened) automatic use response and select another recently-trained use response from the same action neighborhood. Moreover, we predicted that performance on the Stroop-like task would be inversely associated with performance on a task assessing susceptibility to grasp-use conflict. Specifically, better performance on a task requiring inhibition of automatically activated use representations should be associated with more impaired performance in using ‘conflict’ tools, in which use representations must compete with grasp representations. Critically, this relationship should be specific to action; that is, no association is predicted between use of conflict tools and performance on a task requiring inhibition of tool function information.

An additional aim of the investigation was to assess the neural substrates of abnormalities in action competition. We conducted support-vector regression lesion symptom mapping analyses to assess the prediction that a core set of regions, including the left SMG and IFG, would be associated both with reduced competition on the Stroop-like task and poor performance with ‘conflict’ tools (see Goldenberg et al., 2007; Randerath et al., 2010; Watson & Buxbaum, 2015).

2. General methods

Below we report how we determined our sample size, all data exclusions, all inclusion/exclusion criteria, whether inclusion/exclusion criteria were established prior to data analysis, all manipulations, and all measures in our study.

2.1. Participants

We recruited 21 chronic left cerebrovascular accident (LCVA) participants from the Neuro-Cognitive Research Registry at Moss Rehabilitation Research Institute (6 female; mean age = 61.5 years, SD = 10.1, range = 37–80 years; mean education = 14 years, SD = 2.1, range = 11–18 years). Participants were between the ages of 21 and 80, were right handed, and had suffered a single left hemisphere stroke at least 6 months prior to testing. Participants were excluded if they had a history of psychosis, drug or alcohol abuse, co-morbid neurological disorder, or severe language comprehension deficits established with the Western Aphasia Battery (Kertesz, 1982). All participants had previously been tested for with a measure of susceptibility to grasp-use conflict developed in our laboratory (see below), but were not explicitly selected based on their performance. See Table 1 for demographic information for LCVA participants. We also recruited 8 neurotypical participants from the same Research Registry (4 female; mean age = 64.7 years, SD = 9.4, range = 49–80 years; mean education = 15.9 years, SD = 1.7, range = 13–18 years). The neurotypical participants were included to demonstrate that our behavioral paradigm was sufficiently sensitive to generate a typical Stroop-like effect across the experimental conditions, and to assess whether LCVA performance was overall worse than the neurotypical individuals. LCVA participants and neurotypical participants were equivalent in age [t(27) = .79, p = .44; Hedges’ G (HG) effect size, .32], but neurotypical participants, on average, were more highly educated [t(27) = 2.21, p = .04; HG, .89].

Table 1 –

Demographic information and lesion volume for each LCVA participant.

LCVA Participant Age Years of Education Months Post Stroke Gender Lesion Volume (1 mm3)
1 66 12 142 M 69,778
2 60 13 137 M 166,393
3 74 14 116 M 76,301
4 59 14 103 M 82,964
5 80 12 101 F 16,547
6 53 13 145 F 80,020
7 61 12 111 M 47,442
8 65 15 117 M 179,606
9 63 12 86 M 23,141
10 74 14 84 M 64,793
11 54 14 92 M 52,416
12 60 12 68 F 128,897
13 67 14 59 M 51,399
14 37 18 62 M 88,046
15 60 16 59 M 136,565
16 53 12 54 M 92,744
17 74 18 56 F 171,724
18 46 18 47 F 26,714
19 68 14 51 F N/A
20 53 11 37 M 16,977
21 64 16 31 M 5952
Open in a new tab

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.

2.2. Neuroimaging acquisition

Of the twenty-one LCVA individuals who participated in our study, 20 received a research-quality MRI (N = 16) or CT (N = 4) scan. MRI scans included whole-brain T1-weighted MR images collected on a 3T (Siemens Trio, Erlangen, Germany; repetition time = 1620 msec, echo time = 3.87 msec, field of view = 192 × 256 mm, 1 × 1 × 1 mm voxels) or a 1.5T (Siemens Sonata, repetition time = 3,000 msec, echo time = 3.54 msec, field of view = 24 cm, 1.25 × 1.25 × 1.25 mm voxels) scanner, using an eight-channel head coil. Participants for whom MRI scanning was contraindicated underwent whole-brain research CT scans without contrast (60 axial slices, 3–5 mm slice thickness) on a 64-slice Siemens SOMATOM Sensation scanner.

Lesions were manually segmented on each LCVA participant’s high-resolution T1-weighted structural images. Lesioned voxels, consisting of both grey and white matter, were assigned a value of 1 and preserved voxels were assigned a value of 0. Binarized lesion masks were then registered to a standard template (Montreal Neurological Institute “Colin27”) using a symmetric diffeomorphic registration algorithm (Avants, Epstein, Grossman, & Gee, 2008, www.picsl.upenn. edu/ANTS). Volumes were first registered to an intermediate template comprised of healthy brain images acquired on the same scanner. Volumes were then mapped onto the “Colin27” template to complete the transformation into standardized space. To ensure accuracy during the transformation process, lesion maps were subsequently inspected by a neurologist (H.B. Coslett), who was naïve to the behavioral data of the study. Research CT scans were drawn directly onto the “Colin27” template by the same neurologist using MRIcron (https://www.nitrc.org/projects/mricron). For increased accuracy, the pitch of the template was rotated to approximate the slice plane of each LCVA participant’s scan. This method has been demonstrated to achieve high intra- and inter-rater reliability (e.g., see Schnur et al., 2009).

2.3. Experimental task

The experimental Stroop-like task included 12 colored photographs of common tools, selected from a larger normed set of tools and manipulable artifacts from the Bank of Standardized Stimuli (BOSS) database (Brodeur, Dionne-Dostie, Montreuil, & Lepage, 2010; see also; Watson & Buxbaum, 2015). Stimuli are available at https://sites.google.com/site/bosstimuli/. Based on multidimensional scaling and principal components analyses performed by Watson and Buxbaum (2014), a two-dimensional neighborhood space with “amount of arm movement” and “hand posture”2 was derived, and we created 2 pairs of tools selected from that multidimensional space to be close neighbors along the hand posture dimension (corkscrew, wrench; lightbulb, key; see Watson & Buxbaum, 2014). We also created two pairs of tools that were close “function” neighbors (i.e., shared a similar purpose of use; axe, saw; matches, lighter), and two pairs that were unrelated in either action or function (fork, nail clipper; bottle opener, flyswatter).

We verified that action neighbors were similar in hand posture using ratings collected as part of a larger study measuring action similarity and tool use (see Watson & Buxbaum, 2014). Action neighbors were rated to be more similar in the hand posture required for typical use (M, 46; SD,8.49) than function neighbors (M, 13; SD, 4.25) and unrelated items (M, 9; SD, 1.41). Watson and Buxbaum (2014) also derived a measure of semantic similarity between tools, defined as the degree to which two tools are members of the same category. Action neighbors (M, 1.5; SD, .35) and unrelated tools (M, 1.67; SD, .47) were on average less semantically similar than function neighbors (M, 4.75; SD, 0). Finally, tools were balanced for visual similarity: action neighbors (M, 1.38; SD, .18) were approximately equivalent in visual similarity to function neighbors (M, 1.75; SD, .35) and unrelated tools (M, 1.67; SD, .47). See Supplemental Table 1 for rating scores as a function of neighborhood type.

We also ensured that items were balanced for familiarity, visual complexity, and name agreement using normed values derived from the BOSS database (Brodeur et al., 2010). Critically, items were also balanced for manipulability (action neighbors, M, 3.3; SD, .37; function neighbors, M, 3.9; SD, .12; unrelated tools, M, 3.63; SD, .34), defined as the ease with which one could pantomime tool use such that a person could recognize the action. Thus, the gestures associated with items in an action neighborhood were not more difficult to pantomime or more ambiguous in the hand postures required to pantomime tool use. See Supplemental Table 2 for rating scores of each property.

Participants began each block with a two-portion training sequence, after which they completed the experimental task. During the first portion of training, participants were familiarized with one of the tools to appear in the experimental task (e.g., wrench and corkscrew), then shown a 400-by-400 pixel color photograph of each tool on a computer screen. They were instructed to practice pantomiming either by producing the gesture appropriate to that tool (congruent blocks) or to the other tool in the block (incongruent blocks). All responses were made with the participants’ left hands due to the possibility of right hemiparesis. The experimenter scored each gesture in real time, and upon commission of an error, provided verbal and tactile corrective feedback. Practice sequences were repeated until participants achieved a criterion of scoring 3 of 4 trials correct; upon reaching this criterion, the second portion of training began.

In the second portion of training, participants practiced a “Go/No-go” task to further familiarize them with the experimental task and stimulus sequence, and to ensure that pantomime actions were not initiated until participants were cued. Prior to the start of each trial, participants were instructed to keep their left hand relaxed and to depress a button on an E-Prime Stimulus Response Button Box; each trial started with the presentation of a 500-msec fixation cross, followed by presentation of a 400-by-400 pixel color photograph of a tool (a member of the same tool pair shown during the first portion of training), for 4000 msec. Participants kept the button depressed until ready to gesture, whereupon they produced a pantomime and returned their finger to the button. Participants were also instructed that when an animal was presented they should not lift their finger (i.e., “no-go” trials). Each practice sequence consisted of 5 trials which included an animal image and two tools presented twice. Sequences were repeated until participants performed without error.

Following the two training sequences, the experimental sequence was conducted in two sessions, each containing 6 blocks of 28 trials (24 tool trials and 4 no-go trials) presented in random order. Fig. 1 provides a schematic of the trial structure and stimuli. In each block, 2 different tools were presented, and participants gestured either to a depicted tool (congruent blocks) or to the other tool in the block (incongruent blocks). Blocks of trials relevant to a specific type of neighborhood relation [action (A), function (F), unrelated (U)] were presented in AAFFUU-UUFFAA order for half of participants, which was reversed for the other half of participants. All participants took part in block 2 approximately one week after completing block 1. The first block of each neighborhood type was always congruent, and the second block of each neighborhood type was always incongruent. The particular tool pair shown in block one or block two was randomized across participants. Each experimental trial began with the participant depressing a button; a fixation cross was then presented for 500 msec, followed by the presentation of a tool image for 4000 msec. Participants responded by producing a gesture and then returning their finger to the button within the 4000-msec window. Across both sessions, there were 288 experimental trials (144 congruent and 144 incongruent, of which 48 each were action, function, or unrelated), and 48 no-go trials.3 The stimulus presentation code can be found at https://github.com/frankgarcea/GarceaStollBuxbaum.

Fig. 1 –

Fig. 1 –

Open in a new tab

Schematic of trial structure for congruent and incongruent trials in the experimental task. Each trial began with a fixation cross presented for 500 msec, followed by an image of a tool presented for 4000 msec. On congruent trials participants responded by pantomiming the use of the visually presented tool; on incongruent trials participants responded by pantomiming the use of the other tool in the block.

2.4. Test of grasp-use conflict

The test of grasp-use conflict included 40 photographs of manipulable objects (tools) taken from the BOSS database as described above. Tools included items with distinct use actions, including construction tools (e.g., wrench), household articles (e.g., teapot), office supplies (e.g., scissors), and bathroom items (e.g., razor). Half of the items had conflicting actions for use versus grasp (i.e., conflict items; e.g., corkscrew), while half of the items had a single action for both use and grasp (i.e., non-conflict items; e.g., hammer). Tools differed significantly in conflict rating [t(38) = 10.00, p < .001; HG, 3.10], but were equivalent in affordance strength [t(38) = 1.50, p = .14; HG, .45], name agreement [t(38) = 1.13, p = .26; HG, .35], and familiarity [t(38) = 1.50, p = .13; HG, .45; for details of ratings, see Watson & Buxbaum, 2015].

Ten of the 40 tools used in the grasp-use conflict task were also included in the experimental Stroop-like task. LCVA volunteers first performed the test of grasp-use conflict, and were recruited to take part in the experimental task several months later (mean time difference, 20.6 months; SD, 13.7 months), thus mitigating the possibility of practice effects. Although we gave performance feedback during practice trials, no performance feedback was provided for either task during the experimental and grasp-use conflict tasks.

Each trial of the grasp-use conflict task began with the presentation of a 400-by-400 pixel color photograph of a tool on a computer monitor. Participants were asked to “show how you would use the tool as if you were holding and using it” with the left hand. Four practice trials with feedback (using items different than in the task itself) were given at the start of the task. As per Rothi et al. (1991), if a participant gestured the action as if their hand was the tool (body-part-as-object error), they were reminded to “show how you would use the tool as if you were actually holding it in your hand”. The first of these errors was corrected and the participant was permitted a second try.

2.5. Coding of action data

The experimental task and grasp-use conflict task were recorded with a digital camera and scored offline by two trained, reliable coders (Cohen’s Kappa score = 94%) who also demonstrated inter-rater reliability with previous coders in the Buxbaum lab (Cohen’s Kappa >85%; see e.g., Buxbaum, Kyle, & Menon, 2005). Both tasks were coded using a portion of the detailed praxis scoring guidelines used in our previous work (see Buxbaum et al., 2005; Buxbaum, Giovannetti, & Libon, 2000; Watson & Buxbaum, 2015). First, each gesture was given credit for semantic content unless a participant performed a recognizable gesture appropriate for a semantically-related tool, including, in the case of the experimental task, the tool use gesture associated with the paired neighbor. Only gestures that were given credit for semantic content were scored on the spatiotemporal hand action dimension. A hand action error was assigned if the shape or movement trajectory of the hand and/or wrist was flagrantly incorrect, or if the hand or finger was used as part of the tool (i.e., body-part-as-object error, Buxbaum et al., 2005; Watson & Buxbaum, 2015).

2.6. Data analysis approach

Hand action scores from the experimental task were analyzed at the trial level through mixed-effect logistic regression using the R statistical programming environment version 3.5.0 with the lme4 package (Bates, Maechler, Bolker, & Walker, 2015). The R analysis code can be viewed at https://github.com/frankgarcea/GarceaStollBuxbaum. There were three main effects in our analysis: Group (LCVA participants, neurotypical participants), Neighbor Type (action, function, unrelated) and Congruence (congruent, incongruent). Prior to running each model, we assessed if there was an effect of the order of block presentation or number of years of education; we also included random intercepts for subject and item. All mixed-effect models were assessed by mapping the log likelihood ratio of a full and reduced model using a chi square distribution. We removed interaction terms and assessed the full model when testing for significant main effects. An alpha threshold of .05 was used to determine statistical significance, and all effects are reported as log odds.

Performance on the conflict versus non-conflict trials of the test of grasp-use conflict were assessed with t-tests. These scores were also used to derive a grasp-use conflict score, which represents hand action (HA) performance when pantomiming the use of conflict tools relative to non-conflict tools, controlling for overall performance accuracy (HAConflict – HA Non-conflict)/(HA Conflict + HA Non-conflict). Negative-going grasp-use conflict scores indicate worse performance with conflict tools relative to non-conflict tools (i.e., high susceptibility to grasp-use conflict).

A final mixed-effects model assessed how competition in the test of grasp-use conflict related to performance in the experimental task. Grasp-use conflict scores and performance for incongruent action, function, and unrelated neighbor trials in the experimental task were entered as the principal dependent variables, and performance on congruent trials of the relevant condition of the experimental task served as covariates.

2.7. Support vector regression-lesion symptom mapping (SVR-LSM) analyses

SVR-LSM was performed in MATLAB using a toolbox developed by DeMarco and Turkeltaub (2018) (https://github.com/atdemarco/svrlsmgui/). SVR-LSM is a multivariate technique that uses machine learning to determine the association between lesioned voxels and behavior when considering the lesion status of all voxels submitted to the analysis. It overcomes several limitations of voxel-based lesion symptom mapping (VLSM), including inflated false positives from correlated neighboring voxels (Pustina, Avants, Faseyitan, Medaglia, & Coslett, 2018), Type II error due to correction for multiple comparisons (Bennett, Wolford, & Miller, 2009), and uneven statistical power due to biased lesion frequency as a function of vascular anatomy (Mah, Husain, Rees, & Nachev, 2014; Sperber & Karnath, 2017). SVR-LSM has been shown to be superior to VLSM when multiple brain areas are involved in a single behavior (Herbet, Lafargue, & Duffau, 2015; Mah et al., 2014; Mirman, Zhang, Wang, Coslett, & Schwartz, 2015; for discussion, see; Zhang, Kimberg, Coslett, Schwartz, & Wang, 2014).

We performed two SVR-LSM analyses. The first analysis tested the prediction that a reduced incongruence effect in the experimental task would be associated with lesions to the left insula, SMG, IFG, and SLF. The dependent measure was performance on incongruent action neighbor trials, residualized against performance on congruent action neighbor trials. Positive-going residual values indicate better performance on incongruent action neighbor trials. A second SVR-LSM analysis was performed with grasp-use conflict scores, in which hand posture accuracy for conflict tools was residualized against non-conflict hand posture accuracy scores. Positive-going residual values indicate a higher susceptibility to grasp-use conflict when gesturing the use of conflict tools relative to non-conflict tools, which should also be associated with lesions to the left insula, SMG, IFG, and SLF (c.f., Watson & Buxbaum, 2015).

Only voxels lesioned in at least 15% of participants (3 participants) were included. Voxelwise statistical significance was determined using a Monte Carlo style permutation analysis (10,000 iterations) in which the critical behavioral data were randomly assigned to a lesion map, and the resulting map was set to a threshold of p < .05 to determine chance-level likelihood of a lesion-symptom relationship. We then further restricted the resulting map by eliminating any clusters with fewer than 500 contiguous voxels (see Lacey, Skipper-Kallal, Xing, Fama, & Turkeltaub, 2017; Skipper-Kallal, Lacey, Xing, & Turkeltaub, 2017).

The Johns-Hopkins DTI-based probabilistic white matter tractography atlas (Mori et al., 2008) was used to assess the overlap of significant voxels in the SVR-LSM analyses with major white matter fibers (e.g., superior longitudinal fasciculus) at a probability threshold of 25% (see Watson & Buxbaum, 2015).

3. Results

3.1. Excluded and missing trials

Experimental task trials were excluded from analysis if participants deviated from task instructions (e.g., misuse of the button box), or produced a semantic content error. For neurotypical participants, none of the trials were excluded for deviation from task instructions, and .6% of trials (2 trials) were excluded for semantic content errors. Two percent (125 trials) of LCVA participants’ trials were discarded due to deviation from task instructions, and 1% of trials (68 trials) were discarded due to semantic content errors; of the remaining trials, 80% (4728 trials) were correct, 15% of trials (828 trials) were hand action errors, and 5% (299 trials) were other error types (arm, amplitude) not under consideration in this study.

3.2. Experimental task

Fig. 2 and Table 2 present results of the experimental task. We analyzed the LCVA and neurotypical data in two separate models. The first (Model 1) contained the main effect of Group and Congruence along with a Group-by-Congruence interaction term to assess whether the groups differed in congruence effects, collapsing across neighbor type. The second (Model 2) contained the main effects of Group and Neighbor Type along with a Group-by-Neighbor Type interaction term to assess whether the groups differed in performance for different neighbor types, collapsing across congruence (a model containing an interaction term for Group-by-Neighbor Type-by-Congruence failed to converge). A third model (Model 3) contained only the LCVA data, and assessed the relation between Congruence and Neighbor Type.

Fig. 2 –

Fig. 2 –

Open in a new tab

Accuracy in the congruent and incongruent conditions with action, function, and unrelated tool neighbors in the experimental gestural Stroop-like task. Average accuracy for congruent and incongruent trials for A) LCVA participants; B) neurotypical participants. Note that error bars represent standard error of the mean (calculated across participants). Abbreviations. Con., congruent; Incon., incongruent.

Table 2 –

Average accuracy and standard error of the mean for LCVA participants and neurotypical participants.

Group Action Function Unrelated
Congruent Incongruent Congruent Incongruent Congruent Incongruent
Neurotypical .99 (.01) .94 (.03) .99 (.01) .93 (.03) .98 (.01) .92 (.04)
LCVA .80 (.04) .74 (.05) .93 (.03) .92 (.03) .92 (.03) .84 (.04)
Group Differences .19 .20 .06 .01 .06 .08
Open in a new tab

Prior to running these models we evaluated whether the order of block presentation or number of years of education affected performance across the experimental conditions. Whereas order of block presentation had no effect on performance across the LCVA and neurotypical group (all p-values > .70), there was a moderate effect of education on gesture performance in Model 1 [χ2 (1) = 2.93, p = .09] and Model 2 [χ2 (1) = 2.96, p = .09]; thus, we entered number of years of education as a co-variate in Model 1 and 2. Critically, all of the effects observed without controlling for years of education are preserved when controlling for years of education. Finally, prior to conducting Model 3 and the follow-up analysis assessing performance across our experimental and grasp-use conflict tasks, we verified that gesture performance was not modulated by order of block presentation and number of months since stroke by entering both factors as co-variates. Critically, the results remain qualitatively similar when we control for these additional variables.

In Model 1 there was an effect of Congruence [χ2(1) = 60.75, p < .001, −.59 ± .08], indicating that participants were more accurate with congruent trials than incongruent trials. There was also an interaction between Group and Congruence [χ2 (1) = 25.37, p < .001], indicating that neuro-typical participants were more accurate than LCVA participants on congruent trials [χ2 (1) = 13.61, p < .001, 2.65 ± .6], and slightly more accurate on incongruent trials [χ2 (1) = 5.97, p =.01, 1.59 ± .63; see Fig. 2].

In Model 2 there was an effect of Neighbor Type [χ2(2) = 10.38, p < .01]: Action trials were less accurate than function trials [χ2 (1) = 8.05, p < .005, −1.35 ± .36] and unrelated trials [χ2 (1) = 8.9, p < .005, −.76 ± .19], and there was no difference between function and unrelated trials [χ2 (1) = 1.76, p = .18, −.58 ± .4]. Importantly, the Group-by-Neighbor Type interaction was also significant [χ2 (2) = 36.24, p < .001], indicating that neurotypical participants were more accurate than LCVA participants on action neighbor trials (χ2 (1) = 12.84, p < .001, 3.32 ± .09), but were not different on function neighbor trials [χ2 (1) = .04, p = .84, .46 ± 2.25] or unrelated neighbor trials [χ2 (1) = 1.13, p = .29, 1.43 ± 1.34]. Thus, LCVA participants were deficient in their performance with action neighbors only (see Fig. 2).

In Model 3 there was a significant effect of Congruence [χ2(1) = 31.9, p < .001, −.046 ± .08], with better performance on congruent trials than incongruent trials, and a significant effect of Neighbor Type, [χ2 (2) = 14.83, p < .001], indicating that action neighbor trials were less accurate than function [χ2(1) = 11.63, p < .001, −1.47 ± .27] and unrelated neighbor trials [χ2 (1) = 12.15, p < .001, −.9 ± .17]. There was no difference between function neighbors and unrelated neighbors [χ2(1) = 2.4, p = .12, −.61 ± .35].

Finally, there was an interaction between Neighbor Type and Congruence [χ2 (2) = 8.36, p < .05], indicating that the congruence effect for unrelated neighbors was significantly larger in relation to the congruence effect for action [χ2 (1) = 5.36, p < .05, .44 ± .19] and function neighbors [χ2 (1) = 8.95, p < .01, −.73 ± .24]; there was no difference between congruence effects for function and action neighbors [χ2 (1) = .4, p = .53, −.13 ± .21; see Fig. 2].4

3.3. Relationship between experimental and grasp-use conflict tasks

Replicating the results from Watson and Buxbaum (2015), data from the test of grasp-use conflict indicated lower accuracy (i.e., more hand action errors) on conflict (M = .69; SD = .11) relative to non-conflict items [M = .83; SD = .12; t(20) = 6.10, p < .001; HG = 1.17]. In the mixed effect model assessing the relationship between the experimental and grasp-use conflict tasks, there was an interaction between Grasp-Use Conflict Scores and Neighbor Type [χ2 (2) = 19.95, p < .001], indicating that the relationship between grasp-use conflict and incongruent action neighbor performance was stronger than the relationship between grasp-use conflict and performance on incongruent function neighbor trials [χ2 (1) = 13.99, p < .001, 4.64 ± 1.23] and incongruent unrelated trials [χ2 (1) = 6.1, p < .05, 2.21 ± .89; see Fig. 3]. There was no difference between the relationship of grasp-use conflict scores and incongruent function neighbor versus incongruent unrelated trials [χ2(1) = .7, p = .4, 1.21 ± 1.42; see Fig. 3]. Thus, increased conflict between grasp and use is specifically associated with better performance on incongruent action neighbor trials relative to function and unrelated neighbor types in the experimental task.

Fig. 3 –

Fig. 3 –

Open in a new tab

Grasp-use conflict scores plotted against incongruent trials in the experimental gestural Stroop task. Individual LCVA participant data plotted as circles. X-axis: Grasp-Use conflict score. Negative scores indicate higher susceptibility to grasp-use conflict. Y-axis: Incongruent condition of experimental gestural Stroop-like task (% correct).

3.4. Support vector regression-lesion symptom mapping results

Fig. 4 depicts lesion overlap among the 20 participants with high resolution CT or MRI anatomical data.

Fig. 4 –

Fig. 4 –

Open in a new tab

Voxelwise lesion overlap among 20 LCVA participants. Only voxels with at least three participants are included.

Two SVR-LSM analyses were conducted; in the first analysis we identified voxels in which lesions were associated with reduced competition on incongruent action neighbor trials (i.e., hand action performance on incongruent action neighbor trials residualized against hand action performance on congruent action neighbor trials). In the second analysis we identified voxels in which lesions were associated with increased susceptibility to grasp-use competition (i.e., hand action performance when gesturing the use of conflict items residualized against hand action performance with nonconflict items) in the test of grasp-use conflict.

As shown in Fig. 5A, lesions to the left SMG, and somatosensory and motor cortex (pre- and post-central gyri, including ventral premotor cortex) were associated with better performance with incongruent action trials in the experimental task; these lesions extended medially to include the left superior longitudinal fasciculus and the left corona radiata as identified by the JHU white matter atlas (see Table 3; see also Supplemental Figure 1). Next, replicating the voxelwiselesion symptom mapping results from Watson and Buxbaum (2015), we found that lesions to a much larger region of the left IFG, extending medially into the left anterior insula and posteriorly into the left ventral premotor cortex, as well as subcortically to include aspects of the left superior longitudinal fasciculus, were associated with increased susceptibility to grasp-use conflict (See Fig. 5B; see Table 3 for beta values associated with peak MNI coordinates identified in the SVR-LSM analyses).

Fig. 5 –

Fig. 5 –

Open in a new tab

SVR-LSM analyses of data from the experimental gestural Stroop-like task and grasp-use conflict test. A. Voxels associated with reduced incongruence effect (incongruent action trials controlling for congruent action trials; red-to-yellow scale) B. Voxels associated with grasp-use conflict effect (blue-to-green scale). Whole-brain results are rendered in MNI space in increments of 5 mm. SVR-LSM maps are set to a voxelwise threshold of p < .05 with 10,000 iterations of a Monte Carlo style permutation analysis; cluster size >500 contiguous 1 mm3 voxels.

Table 3 –

Peak voxels in MNI space identified in the SVR-LSM analysis of the action neighbor incongruence effect from the experimental Stroop-like task, and the grasp-use conflict score. The overlap between peak voxels and major white matter fibers was verified with the Johns Hopkins DTI-based probabilistic white matter tractography atlas.

Action Neighbor Incongruence Effect
X Y Z Beta value of peak voxel Anatomical size (1 mm3)
Superior Longitudinal Fasciculus −37 −38 22 10.0 18,366
Left Anterior Corona Radiata −28 18 28 9.11 1743
Grasp-Use Conflict Score
Superior Longitudinal Fasciculus −39 −9 26 10.0 18,050
Open in a new tab

4. General discussion

This study assessed the hypothesis that reduced automatic activation of use actions in LCVA has specific consequences for patterns of competition within and between tools. We demonstrated that LCVA participants exhibit reduced competition between neighboring tool use actions on a Stroop-like task, as evidenced by a reduced “cost” of producing a gesture associated with a tool’s action neighbor rather than by the depicted tool. Second, we showed that this reduction in competition between tool action neighbors–but not between tool function neighbors–was associated with impaired gesture production in a separate task pitting use against grasp actions within ‘conflict’ tools. Finally, based on SVR-LSM analyses, we also demonstrated that lesions to voxels in the left insula, pre- and post-central gyri, SLF, and SMG were associated with reduced competition between tool action neighbors, and that lesions to a large swath of left IFG and SLF were associated with susceptibility to grasp-use conflict. Together, these data suggest that abnormalities in gesture performance in apraxic individuals are not merely the result of weakened activation of use action representations, but also reflect downstream consequences of weakened activation for competition and action selection. Moreover, they confirm prior observations suggesting that abnormalities in tool action competition are associated with lesions to a specific subset of peri-Sylvian regions in the left hemisphere.

To understand the findings from our experimental gestural Stroop-like task, imagine a scenario in which you are asked to perform a standard Stroop task (name the color of the word, ignore the content of the text), but you are a poor reader and have trouble deriving meaning from written words. If the written word does not readily activate that word’s meaning, the meaning will not strongly interfere with naming the color of the word, even when that color is incongruent with the word meaning. This is analogous to the situation we posit for patients with limb apraxia. Prior work shows that for patients with limb apraxia, automatic action activation when viewing tools is absent or diminished (Lee et al., 2014; Myung et al., 2010). In our study, weakened automatic action activation may be easier to inhibit, rendering it easier for LCVA participants to produce tool-incongruent actions demanded by the Stroop-like task. Now, imagine that one is a fluent reader and readily derives meaning from words. In this case, the meaning of the word interferes with naming the color of the word when they are incongruent. This is analogous to the normal interference we see when neurotypical participants try to produce a task-specified action that is incongruent with that automatically evoked by a viewed tool (see Fig. 2 and Table 2).

To continue the above analogy, assume that now, instead of a Stroop-like task where one has to suppress the words’ meanings in order to name the color, we focus on a word comprehension task with words having two meanings. One meaning is learned very early in life and based closely on the words’ sounds (e.g., an onomatopoeia), and the other is learned later in life and requires more controlled retrieval of semantic knowledge. Losses in semantic knowledge will make it much easier to produce the meaning associated with the words’ sounds than the other meaning. This is analogous to the situation with limb apraxics in which they are able to grasp tools that they have trouble using, and moreover, have more trouble with conflict tools in which grasp activations (analogous to activation of words’ sounds in our example) readily override use activations (analogous to semantic knowledge in our example).

To tie the two tasks together, consider that with weakened automatic use activation, one may find it easier to override that weak activation to produce another action. If that “other action” is what the task calls for, as is the case in a Stroop task, then performance will be benefitted. If, on the other hand, the “other action” is incorrect, as is the case with use of conflict objects, then performance will be negatively influenced.

4.1. The neuroanatomy of action selection and competition

Considerable evidence suggests that the left IFG is critical for performance across a range of semantic control tasks, including naming, word-picture matching, and synonymy judgments with pictures, particularly when items are close neighbors or when demands on control are otherwise high (e.g., Hamilton & Martin, 2005; Mirman & Graziano, 2013; Novick, Kan, Trueswell, & Thompson-Schill, 2009; Nozari, Freund, Breining, Rapp, & Gordon, 2016; Ries, Karzmark, Navarrete, Knight, & Dronkers, 2015; Schnur, Lee, Coslett, Schwartz, & Thompson-Schill, 2005; Schnur et al., 2009; Stampacchia et al., 2018; Thompson-Schill, D’Esposito, Aguirre, & Farah, 1997; Thompson-Schill, D’Esposito, & Kan, 1999; Thompson-Schill et al., 1998). Moreover, research in patients with language deficits demonstrates that competition among neighboring word representations is exacerbated in the presence of lesions to the left IFG, resulting in a greater number of selection-related errors in picture naming tasks (e.g., see Ries et al., 2015; Schnur et al., 2009; Thompson-Schill et al., 1998). Consistent with the proposal that IFG and temporo-parietal regions are nodes of a left hemisphere neural network critical to selection in a broad range of tasks requiring timely and efficient access to semantic information (e.g., Cristescu, Devlin, & Nobre, 2006; Nagel, Schumacher, Goebel, & D’Esposito, 2008; Rodd, Davis, & Johnsrude, 2005)d including action tasksdwe have previously suggested that the left SMG briefly buffers candidate actions during the process of selection, whereas the left IFG biases processing such that irrelevant actions are inhibited and task-relevant actions are selected for subsequent production (Buxbaum, 2017; for a related proposal, see; Cisek, 2007; Cisek & Kalaska, 2010).

The present results indicate that lesions to partly overlapping brain regions in an IFG-IPL circuit can give rise to abnormalities in both within-tool as well as between-tool action competition. Previous work has demonstrated that grasp information is available more rapidly than use information within single tools; this results in slowing when grasp and use fail to converge on a single response (Jax & Buxbaum, 2010; Lee et al., 2013). From the perspective of classic cognitive winner-take-all models, grasp information rapidly evoked by the shape and size of currently perceived tools must be inhibited in order to select tool use representations (e.g., see Jax & Buxbaum, 2010). Moreover, lesions to the left IFG, insula, and SMG have been associated with abnormal within-tool competition resolution due to increased grasp-on-use conflict (e.g., Jax & Buxbaum, 2013; Watson & Buxbaum, 2015). There are two points of divergence with respect to the neuroanatomy identified in the present lesion-symptom mapping analysis. First, we did not observe significant SMG voxels associated with within-tool (Grasp-Use) conflict in the present study, even at a relaxed threshold (see Supplemental Figure 2). Second, performance in our experimental Stroop-like task largely spared the left IFG (see Fig. 5A). Prior research indicates that the left IFG is critical for conflict resolution among competing alternative responses (Schnur et al., 2009). As noted, prior research indicates that there is weak and slowed automatic activation of tool use actions in limb apraxia (Lee et al., 2014; Myung et al., 2010). With respect to the present experiment, we may speculate that when the automatic typical tool use actions are only weakly activated, the highly practiced, recent incongruent actions demanded by the task may more easily “win” the selection process without substantial IFG intervention. In contrast, the left IFG was identified in our test of within-tool conflict (pantomiming the use of a conflict item), where participants were required to inhibit a dominant grasp-based response in favor of a (weakly activated) use-based response. An additional possibility is that these anatomical differences reflect interesting differences between the portions of the IFG-IPL circuit critical for between- versus within-tool competition resolution (but see Watson & Buxbaum, 2015). Future studies will be required to further explore whether the same regions mediate both aspects of competition.

From a computational perspective, there are at least two non-exclusive interactive activation accounts that may explain how reduced activation of tool use actions can lead to reduced Stroop costs (i.e., improved ability to produce target-incongruent actions). One possibility is that less precise activation of tool use actions entails weak activation both of the target action as well as action neighbors, thus facilitating production of the neighbors. Alternatively, weakened automatic activation of use actions may be associated with weaker lateral inhibition of incongruent neighboring actions, rendering them easier to produce. Similar accounts have been proposed in the language domain. For example, a recent computational modeling endeavor suggested that when activation of non-target neighbors is strong, there is a net inhibitory effect on target word selection (in this case, driven in part by lateral inhibition among competing neighbors); in contrast, weak activation of non-target neighbors caused a net facilitative effect on target selection via excitatory connections among the target and non-target neighbors (Chen & Mirman, 2012). On this account, then, weak activation of use representations may lead to net facilitation of the target action relative to its neighbors.

Not all semantic neighbors exert equal effects, as close semantic neighbors will compete to a greater extent than distant semantic neighbors. For example, naming a picture of a dog is frequently slower in the presence of a near semantic neighbor (e.g., cat) than a distant semantic neighbor (e.g., whale) or an unrelated distractor (e.g., couch; see Damian et al., 2001; Fieder et al., 2018; Rose et al., 2018; Vieth et al., 2014; but see Mahon, Costa, Peterson, Vargas, & Caramazza, 2007). The semantic properties that influence neighborhood distance may plausibly include shape, size, color, typical location, or other taxonomic or event-related (thematic) features. In addition, in the tool domain it has become increasingly clear that action is a core semantic feature. Our prior work suggests that tool action similarity takes into account aspects of action appearance that include magnitude of arm movement and posture of the hand (Watson & Buxbaum, 2014), and it is likely that additional features may determine which tools are close in neighborhood space. fMRI experiments in neurotypical adults have reported that tool-related actions may be represented with respect to their visual appearance and motion (Beauchamp, Lee, Haxby, & Martin, 2002; Tucciarelli et al., 2019), whether they are performed with a tool or the hands (Bracci, Cavina-Pratesi, Ietswaart, Caramazza, & Peelen, 2012; Gallivan, McLean, Valyear, & Culham, 2013), whether they are person-directed or object-directed (e.g., see Wurm & Caramazza, 2019) as well as such non-sensory features as the number of sequences they entail (Gallivan, Johnsrude, & Flanagan, 2016) or their abstract function or goal (Wurm & Lingnau, 2015; for review, see; Lingnau & Downing, 2015).

The present data are potentially inconsistent with at least two alternative explanations for action impairments in apraxia following LCVA. First, it has been proposed that technical reasoning, or the ability to derive manipulation information online on the basis of the visual structure of objects, may be the core mechanism disrupted in the apraxia syndrome (e.g., see Osiurak, 2014; Osiurak & Badets, 2016; but see Buxbaum, 2017). Second, it has been argued that apraxics’ apparent deficit in using ‘conflict’ tools is attributable to the greater visual complexity and variability of such objects as compared to non-conflict objects (Bub, Masson, & van Mook, 2018). However, neither of these accounts can accommodate the present data indicating that performance on a Stroop-like gesture task is differentially modulated by neighbor type. Moreover, these accounts cannot explain the observed specific association between reduced interference on the Stroop-like task with action neighbors and impaired performance with conflict tools. Although it is possible that structural affordances, visual complexity, and mechanical problem-solving ability may all influence praxis performance (for discussion, see Buxbaum, 2017; Garcea & Buxbaum, 2019; Garcea & Mahon, 2019), these alternative explanations alone cannot account for the pattern of abnormal within-item and between-item tool use competition observed here.

The implications of grasp-on-use conflict identified in our study may be of relevance for the performance of patients with limb apraxia when using real tools. For example, Sirigu et al. (1995) reported the performance of Case LL, a limb apraxic who was impaired when asked to grasp an object to use it, but was spared when asked to perform simple reach-to-grasp actions (for discussion, see Buxbaum, Sirigu, Schwartz, & Klatzky, 2003). In a similar vein, we may speculate that the LCVA participants in our study should make more hand-action errors with use of real 3 dimensional conflict tools than non-conflict tools. In general, patients with limb apraxic make relatively few action errors when using real tools, likely because the structural properties and somatosensory feedback from real objects provide strong “bottom-up” affordances for how the object should be grasped, thus minimizing the opportunities to detect deficits (e.g., see Osiurak et al., 2008; Randerath, Li, Goldenberg, & Hermsdorfer, 2009). Nevertheless, we might speculate that these affordances are more helpful for non-conflict than conflict tools, because only in the former case do affordances provide information consistent with the use of the tool. This will be an interesting question for future analyses with sensitive methods (e.g., movement kinematics).

There are several limitations of our study worth noting. In our experimental task we used six pairs of tools to study action competition; thus it will be important for future research to identify a larger set of items from which action distance can be derived to study competition among tool action neighbors. Second, although we observed a typical Stroop-like effect for unrelated trials (i.e., better performance with congruent than incongruent items), participants failed to exhibit a typical congruence effect on function trials, a result that was unexpected. Critically however, variability in performance for function neighbors in the experimental task was not related to variability in our measure of grasp-use conflict at the single-item level. Thus, there are competition abnormalities in between-tool actions bearing a specific relationship to competition abnormalities in within-tool actions. On a broader scale, the relatively better performance of LCVA participants on function as compared to action neighborhood trials is consistent with the known dissociation between function and manipulation knowledge in apraxia (Buxbaum & Saffran, 2002; Buxbaum, Veramonti, & Schwartz, 2000; see also Garcea & Mahon, 2012 for dissociations in neurotypical individuals).

5. Conclusion

In conclusion, the present data complement a number of lines of prior evidence indicating that left hemisphere stroke may be characterized by abnormalities in the competitive processes by which task-appropriate tool action representations are selected from among candidate actions. Although the precise nature of these representations is unclear, they need not be motor plans carrying instructions for the programming of specific muscles (c.f. Oostwoud Wijdenes, Ivry, & Bays, 2016; Wong & Haith, 2017). Rather, we suggest that candidate actions compete in prior processing stages, in which visual action appearance (Buxbaum, 2017), abstract kinematic trajectory (Wong, Haith, & Krakauer, 2015), higher level task goals, and other relatively abstract features determine action similarity (and thus, patterns of competition). Rather than merely reflecting a loss of action knowledge, apraxia may be characterized by the consequences of slow and weakened action activation (see Lee et al., 2014) for the downstream selection of appropriate action representations that inform subsequent motor planning.

Supplementary Material

1

Acknowledgements

We would like to thank Cortney Howard, Leyla Tarhan, and Louisa Smith for coding participants’ gestures; Christine Watson for sharing her skills and tool norming data; and H. Branch Coslett for help with lesion segmentation. This research was supported by NIH grant R01 NS099061 to L.J. Buxbaum. Preparation of this manuscript was supported by a Moss Rehabilitation Research Institute/University of Pennsylvania postdoctoral training fellowship (NIH 5T32HD071844-05).

Footnotes

The authors declare no competing financial interests.

Open practices

The study in this article earned an Open Materials badge for transparent practices. Materials and data for the study are available at https://sites.google.com/site/bosstimuli/, https://github.com/frankgarcea/GarceaStollBuxbaum and https://github.com/atdemarco/svrlsmgui/.

Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cortex.2019.05.021.

2

We use the term “hand action” and “hand posture” interchangeably throughout the text.

3

Note that no part of the study procedures or analyses was pre-registered prior to the research being conducted.

4

We ran a two-way ANOVA with factors of Congruence and Neighbor Type to verify that the Neighbor-Type-by-Congruence interaction observed in LCVA participants was not observed in neurotypical adults. The ANOVA yielded a significant main effect of Congruence (F(1,42) = 7.9, p < .01), with higher performance for congruent trials relative to incongruent trials, but no main effect of Neighbor Type nor an interaction between the two factors (both Fs < 1).

REFERENCES

  1. Avants BB, Epstein CL, Grossman M, & Gee JC (2008). Symmetric diffeomorphic image registration with cross-correlation: Evaluating automated labeling of elderly and neurodegenerative brain. Medical Image Analysis, 12(1), 26–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Bates D, Maechler M, Bolker B, & Walker S (2015). Fitting linear mixed-effects models using lme4. Journal of Statistical Software, 67, 1–48. [Google Scholar]
  3. Beauchamp MS, Lee KE, Haxby JV, & Martin A (2002). Parallel visual motion processing streams for manipulable objects and human movements. Neuron, 34(1), 149–159. [DOI] [PubMed] [Google Scholar]
  4. Bennett CM, Wolford GL, & Miller MB (2009). The principled control of false positives in neuroimaging. Social Cognitive and Affective Neuroscience, 4(4), 417–422. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bracci S, Cavina-Pratesi C, Ietswaart M, Caramazza A, & Peelen MV (2012). Closely overlapping responses to tools and hands in left lateral occipitotemporal cortex. Journal of Neurophysiology, 107(5), 1443–1456. 10.1152/jn.00619.2011. [DOI] [PubMed] [Google Scholar]
  6. 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]
  7. Bub DN, & Masson ME (2012). On the dynamics of action representations evoked by names of manipulable objects. Journal of Experimental Psychology General, 141(3), 502–517. 10.1037/a0026748. [DOI] [PubMed] [Google Scholar]
  8. Bub DN, Masson MEJ, & van Mook H (2018). Switching between lift and use grasp actions. Cognition, 174, 28–36. 10.1016/j.cognition.2018.01.013. [DOI] [PubMed] [Google Scholar]
  9. Buxbaum LJ (2017). Learning, remembering, and predicting how to use tools: Distributed neurocognitive mechanisms: Comment on Osiurak and Badets (2016). Psychological Review, 124(3), 346–360. 10.1037/rev0000051. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. 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]
  11. 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]
  12. Buxbaum LJ, & Saffran EM (2002). Knowledge of object manipulation and object function: Dissociations in apraxic and nonapraxic subjects. Brain and Language, 82(2), 179–199. [DOI] [PubMed] [Google Scholar]
  13. Buxbaum LJ, Sirigu A, Schwartz MF, & Klatzky R (2003). Cognitive representations of hand posture in ideomotor apraxia. Neuropsychologia, 41(8), 1091–1113. [DOI] [PubMed] [Google Scholar]
  14. Buxbaum LJ, Veramonti T, & Schwartz MF (2000). Function and manipulation tool knowledge in apraxia: Knowing ‘what for’ but not ‘how’. Neurocase, 6, 83–97. [Google Scholar]
  15. Chen Q, & Mirman D (2012). Competition and cooperation among similar representations: Toward a unified account of facilitative and inhibitory effects of lexical neighbors. Psychological Review, 119(2), 417–430. 10.1037/a0027175. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Cisek P (2007). Cortical mechanisms of action selection: The affordance competition hypothesis. Philosophical Transactions of the Royal Society B Biological Sciences, 362(1485), 1585–1599. 10.1098/rstb.2007.2054. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Cisek P, & Kalaska JF (2005). Neural correlates of reaching decisions in dorsal premotor cortex: Specification of multiple direction choices and final selection of action. Neuron, 45(5), 801–814. 10.1016/j.neuron.2005.01.027. [DOI] [PubMed] [Google Scholar]
  18. Cisek P, & Kalaska JF (2010). Neural mechanisms for interacting with a world full of action choices. Annual Review of Neuroscience, 33, 269–298. 10.1146/annurev.neuro.051508.135409. [DOI] [PubMed] [Google Scholar]
  19. Collins AM, & Loftus EF (1975). A spreading-activation theory of semantic processing. Psychological Review, 82, 407–428. [Google Scholar]
  20. Cristescu TC, Devlin JT, & Nobre AC (2006). Orienting attention to semantic categories. Neuroimage, 33(4), 1178–1187. 10.1016/j.neuroimage.2006.08.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Damian MF, Vigliocco G, & Levelt WJ (2001). Effects of semantic context in the naming of pictures and words. Cognition, 81(3), B77–B86. [DOI] [PubMed] [Google Scholar]
  22. DeMarco AT, & Turkeltaub PE (2018). A multivariate lesion symptom mapping toolbox and examination of lesion-volume biases and correction methods in lesion-symptom mapping. Human Brain Mapping, 39(11), 4169–4182. 10.1002/hbm.24289. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Desimone R, & Duncan J (1995). Neural mechanisms of selective visual attention. Annual Review of Neuroscience, 18, 193–222. 10.1146/annurev.ne.18.030195.001205. [DOI] [PubMed] [Google Scholar]
  24. Fieder N, Wartenburger I, & Abdel Rahman R (2018). A close call: Interference from semantic neighbourhood density and similarity in language production. Memory & Cognition. 10.3758/s13421-018-0856-y. [DOI] [PubMed] [Google Scholar]
  25. Gallivan JP, Johnsrude IS, & Flanagan JR (2016). Planning ahead: Object-directed sequential actions decoded from human frontoparietal and occipitotemporal networks. Cerebral Cortex, 26(2), 708–730. 10.1093/cercor/bhu302. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Gallivan JP, McLean DA, Valyear KF, & Culham JC (2013). Decoding the neural mechanisms of human tool use. Elife, 2, e00425 10.7554/eLife.00425. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Garcea FE, & Buxbaum LJ (2019). Gesturing tool use and tool transport actions modulates inferior parietal functional connectivity with the dorsal and ventral object processing pathways. Human Brain Mapping. 10.1002/hbm.24565. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Garcea FE, & Mahon BZ (2012). What is in a tool concept? Dissociating manipulation knowledge from function knowledge. Memory & Cognition, 40(8), 1303–1313. 10.3758/s13421-012-0236-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Garcea FE, & Mahon BZ (2019). The how and what of object knowledge in the human brain In Schiller NO, & de Zubicaray G (Eds.), The oxford handbook of neurolinguistics. Oxford, UK: Oxford University Press. [Google Scholar]
  30. Goldenberg G, Hermsdorfer J, Glindemann R, Rorden C, & Karnath HO (2007). Pantomime of tool use depends on integrity of left inferior frontal cortex. Cerebral Cortex, 17(12), 2769–2776. 10.1093/cercor/bhm004. [DOI] [PubMed] [Google Scholar]
  31. Hamilton AC, & Martin RC (2005). Dissociations among tasks involving inhibition: A single-case study. Cognitive, Affective & Behavioral Neuroscience, 5(1), 1–13. [DOI] [PubMed] [Google Scholar]
  32. Herbet G, Lafargue G, & Duffau H (2015). Rethinking voxel-wise lesion-deficit analysis: A new challenge for computational neuropsychology. Cortex, 64, 413–416. 10.1016/j.cortex.2014.10.021. [DOI] [PubMed] [Google Scholar]
  33. Humphreys GW (1998). Neural representation of objects in space: A dual coding account. Philosophical Transactions of the Royal Society of London Series B Biological Sciences, 353(1373), 1341–1351. 10.1098/rstb.1998.0288. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Jax SA, & Buxbaum LJ (2010). Response interference between functional and structural actions linked to the same familiar object. Cognition, 115(2), 350–355. 10.1016/j.cognition.2010.01.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Jax SA, & Buxbaum LJ (2013). Response interference between functional and structural object-related actions is increased in patients with ideomotor apraxia. Journal of Neuropsychology, 7(1), 12–18. 10.1111/j.1748-6653.2012.02031.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Johnson-Frey SH (2004). The neural bases of complex tool use in humans. Trends in Cognitive Sciences, 8(2), 71–78. 10.1016/j.tics.2003.12.002. [DOI] [PubMed] [Google Scholar]
  37. Kan IP, & Thompson-Schill SL (2004). Selection from perceptual and conceptual representations. Cognitive, Affective & Behavioral Neuroscience, 4(4), 466–482. [DOI] [PubMed] [Google Scholar]
  38. Kertesz A (1982). Western Aphasia Battery. New York: Grune &Stratton. [Google Scholar]
  39. Kimberg DY, & Farah MJ (1993). A unified account of cognitive impairments following frontal lobe damage: The role of working memory in complex, organized behavior. Journal of Experimental Psychology General, 122(4), 411–428. [DOI] [PubMed] [Google Scholar]
  40. Kim JN, & Shadlen MN (1999). Neural correlates of a decision in the dorsolateral prefrontal cortex of the macaque. Nature Neuroscience, 2(2), 176–185. 10.1038/5739. [DOI] [PubMed] [Google Scholar]
  41. Lacey EH, Skipper-Kallal LM, Xing S, Fama ME, & Turkeltaub PE (2017). Mapping common aphasia assessments to underlying cognitive processes and their neural substrates. Neurorehabilitation and Neural Repair, 31(5), 442–450. 10.1177/1545968316688797. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Lee CL, Middleton E, Mirman D, Kalenine S, & Buxbaum LJ (2013). Incidental and context-responsive activation of structure- and function-based action features during object identification. Journal of Experimental Psychology Human Perception and Performance, 39(1), 257–270. 10.1037/a0027533. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Lee CL, 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]
  44. Levelt WJ, Roelofs A, & Meyer AS (1999). A theory of lexical access in speech production. The Behavioral and Brain Sciences, 22(1), 1–38. discussion 38–75. [DOI] [PubMed] [Google Scholar]
  45. Lingnau A, & Downing PE (2015). The lateral occipitotemporal cortex in action. Trends in Cognitive Sciences, 19(5), 268–277. 10.1016/j.tics.2015.03.006. [DOI] [PubMed] [Google Scholar]
  46. Luce PA, & Pisoni DB (1998). Recognizing spoken words: The neighborhood activation model. Ear and Hearing, 19(1), 1–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Magnuson JS, Dixon JA, Tanenhaus MK, & Aslin RN (2007). The dynamics of lexical competition during spoken word recognition. Cognitive Science, 31(1), 133–156. 10.1080/03640210709336987. [DOI] [PubMed] [Google Scholar]
  48. Mah YH, Husain M, Rees G, & Nachev P (2014). Human brain lesion-deficit inference remapped. Brain, 137(Pt 9), 2522–2531. 10.1093/brain/awu164. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Mahon BZ, Costa A, Peterson R, Vargas KA, &Caramazza A (2007). Lexical selection is not by competition: A reinterpretation of semantic interference and facilitation effects in the picture-word interference paradigm. Journal of Experimental Psychology Learning, Memory, and Cognition, 33(3), 503–535. 10.1037/0278-7393.33.3.503. [DOI] [PubMed] [Google Scholar]
  50. Martin A (2016). GRAPES-Grounding representations in action, perception, and emotion systems: How object properties and categories are represented in the human brain. Psychonomic Bulletin & Review, 23(4), 979–990. 10.3758/s13423-015-0842-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. McKinstry C, Dale R, & Spivey MJ (2008). Action dynamics reveal parallel competition in decision making. Psychological Science, 19(1), 22–24. 10.1111/j.1467-9280.2008.02041.x. [DOI] [PubMed] [Google Scholar]
  52. Mirman D, & Graziano KM (2013). The neural basis of inhibitory effects of semantic and phonological neighbors in spoken word production. Journal of Cognitive Neuroscience, 25(9), 1504–1516. 10.1162/jocn_a_00408. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Mirman D, & Magnuson JS (2008). Attractor dynamics and semantic neighborhood density: Processing is slowed by near neighbors and speeded by distant neighbors. Journal of Experimental Psychology Learning, Memory, and Cognition, 34(1), 65–79. 10.1037/0278-7393.34.1.65. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. 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]
  55. Mori S, Oishi K, Jiang H, Jiang L, Li X, Akhter K, et al. (2008).Stereotaxic white matter atlas based on diffusion tensor imaging in an ICBM template. Neuroimage, 40(2), 570–582. 10.1016/j.neuroimage.2007.12.035. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Myung JY, Blumstein SE, & Sedivy JC (2006). Playing on the typewriter, typing on the piano: Manipulation knowledge of objects. Cognition, 98(3), 223–243. 10.1016/j.cognition.2004.11.010. [DOI] [PubMed] [Google Scholar]
  57. Myung JY, 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]
  58. Nagel IE, Schumacher EH, Goebel R, & D’Esposito M (2008). Functional MRI investigation of verbal selection mechanisms in lateral prefrontal cortex. Neuroimage, 43(4), 801–807. 10.1016/j.neuroimage.2008.07.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Novick JM, Kan IP, Trueswell JC, & Thompson-Schill SL (2009). A case for conflict across multiple domains: Memory and language impairments following damage to ventrolateral prefrontal cortex. Cognitive Neuropsychology, 26(6), 527–567. 10.1080/02643290903519367. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Nozari N, Freund M, Breining B, Rapp B, & Gordon B (2016). Cognitive control during selection and repair in word production. Language, Cognition and Neuroscience, 31(7), 886–903. 10.1080/23273798.2016.1157194. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Oostwoud Wijdenes L, Ivry RB, & Bays PM (2016). Competition between movement plans increases motor variability: Evidence of a shared resource for movement planning. Journal of Neurophysiology, 116(3), 1295–1303. 10.1152/jn.00113.2016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Osiurak F (2014). What neuropsychology tells us about human tool use? The four constraints theory (4CT): Mechanics, space, time, and effort. Neuropsychology Review, 24(2), 88–115. 10.1007/s11065-014-9260-y. [DOI] [PubMed] [Google Scholar]
  63. Osiurak F, Aubin G, Allain P, Jarry C, Richard I, & Le Gall D (2008). Object utilization and object usage: A single-case study. Neurocase, 14(2), 169–183. 10.1080/13554790802108372. [DOI] [PubMed] [Google Scholar]
  64. Osiurak F, & Badets A (2016). Tool use and affordance:Manipulation-based versus reasoning-based approaches. Psychological Review, 123(5), 534–568. 10.1037/rev0000027. [DOI] [PubMed] [Google Scholar]
  65. Pastor-Bernier A, & Cisek P (2011). Making choices between rules or between actions. Neuron, 70(3), 382–384. 10.1016/j.neuron.2011.04.017. [DOI] [PubMed] [Google Scholar]
  66. Pustina D, Avants B, Faseyitan OK, Medaglia JD, &Coslett HB (2018). Improved accuracy of lesion to symptom mapping with multivariate sparse canonical correlations. Neuropsychologia, 115, 154–166. 10.1016/j.neuropsychologia.2017.08.027. [DOI] [PubMed] [Google Scholar]
  67. Randerath J, Goldenberg G, Spijkers W, Li Y, &Hermsdorfer J (2010). Different left brain regions are essential for grasping a tool compared with its subsequent use. Neuroimage, 53(1), 171–180. 10.1016/j.neuroimage.2010.06.038. [DOI] [PubMed] [Google Scholar]
  68. Randerath J, Li Y, Goldenberg G, & Hermsdorfer J (2009).Grasping tools: Effects of task and apraxia. Neuropsychologia, 47(2), 497–505. 10.1016/j.neuropsychologia.2008.10.005. [DOI] [PubMed] [Google Scholar]
  69. Ries SK, Karzmark CR, Navarrete E, Knight RT, &Dronkers NF (2015). Specifying the role of the left prefrontal cortex in word selection. Brain and Language, 149, 135–147. 10.1016/j.bandl.2015.07.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  70. Rodd JM, Davis MH, & Johnsrude IS (2005). The neural mechanisms of speech comprehension: fMRI studies of semantic ambiguity. Cerebral Cortex, 15(8), 1261–1269. 10.1093/cercor/bhi009. [DOI] [PubMed] [Google Scholar]
  71. Rose SB, Aristei S, Melinger A, & Abdel Rahman R (2018). The closer they are, the more they interfere: Semantic similarity of word distractors increases competition in language production. Journal of Experimental Psychology Learning, Memory, and Cognition. 10.1037/xlm0000592. [DOI] [PubMed] [Google Scholar]
  72. Rothi LJ, Ochipa C, & Heilman KM (1991). A cognitive neuropsychological model of limb praxis. Cognitive Neuropsychology, 8(6), 443–458. [Google Scholar]
  73. Rounis E, & Humphreys G (2015). Limb apraxia and the “affordance competition hypothesis”. Frontiers in Human Neuroscience, 9, 429 10.3389/fnhum.2015.00429. [DOI] [PMC free article] [PubMed] [Google Scholar]
  74. Schnur TT, Lee E, Coslett HB, Schwartz MF, & Thompson-Schill SL (2005). When lexical selection gets tough, the LIFG gets going: A lesion analysis study of interference during word production. Brain and Language, 95(1), 12–13. 10.1016/j.bandl.2005.07.008. [DOI] [Google Scholar]
  75. Schnur TT, Schwartz MF, Kimberg DY, Hirshorn E,Coslett HB, & Thompson-Schill SL (2009). Localizing interference during naming: Convergent neuroimaging and neuropsychological evidence for the function of Broca’s area. Proceedings of the National Academy of Sciences of the United States of America, 106(1), 322–327. 10.1073/pnas.0805874106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  76. Sirigu A, Cohen L, Duhamel JR, Pillon B, Dubois B, & Agid Y (1995). A selective impairment of hand posture for object utilization in apraxia. Cortex, 31(1), 41–55. [DOI] [PubMed] [Google Scholar]
  77. Skipper-Kallal LM, Lacey EH, Xing S, & Turkeltaub PE (2017). Functional activation independently contributes to naming ability and relates to lesion site in post-stroke aphasia. Human Brain Mapping, 38(4), 2051–2066. 10.1002/hbm.23504. [DOI] [PMC free article] [PubMed] [Google Scholar]
  78. Sperber C, & Karnath HO (2017). Impact of correction factors in human brain lesion-behavior inference. Human Brain Mapping, 38(3), 1692–1701. 10.1002/hbm.23490. [DOI] [PMC free article] [PubMed] [Google Scholar]
  79. Stampacchia S, Thompson HE, Ball E, Nathaniel U,Hallam G, Smallwood J, et al. (2018). Shared processes resolve competition within and between episodic and semantic memory: Evidence from patients with LIFG lesions. Cortex, 108, 127–143. 10.1016/j.cortex.2018.07.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  80. Thompson-Schill SL, D’Esposito M, Aguirre GK, & Farah MJ (1997). Role of left inferior prefrontal cortex in retrieval of semantic knowledge: A reevaluation. Proceedings of the National Academy of Sciences of the United States of America, 94(26), 14792–14797. [DOI] [PMC free article] [PubMed] [Google Scholar]
  81. Thompson-Schill SL, D’Esposito M, & Kan IP (1999). Effects of repetition and competition on activity in left prefrontal cortex during word generation. Neuron, 23(3), 513–522. [DOI] [PubMed] [Google Scholar]
  82. Thompson-Schill SL, Swick D, Farah MJ, D’Esposito M, Kan IP, & Knight RT (1998). Verb generation in patients with focal frontal lesions: A neuropsychological test of neuroimaging findings. Proceedings of the National Academy of Sciences of the United States of America, 95(26), 15855–15860. [DOI] [PMC free article] [PubMed] [Google Scholar]
  83. Tucciarelli R, Wurm MF, Baccolo E, & Lingnau A (2019). The representational space of observed actions. BioRxiv, 1–26. 10.1101/592071. [DOI] [PMC free article] [PubMed] [Google Scholar]
  84. Vieth HE, McMahon KL, & de Zubicaray GI (2014). Feature overlap slows lexical selection: Evidence from the picture-word interference paradigm. The Quarterly Journal of Experimental Psychology (Hove), 67(12), 2325–2339. 10.1080/17470218.2014.923922. [DOI] [PubMed] [Google Scholar]
  85. Vigliocco G, Vinson DP, Lewis W, & Garrett MF (2004). Representing the meanings of object and action words: The featural and unitary semantic space hypothesis. Cognitive Psychology, 48(4), 422–488. 10.1016/j.cogpsych.2003.09.001. [DOI] [PubMed] [Google Scholar]
  86. Watson CE, & Buxbaum LJ (2014). Uncovering the architecture of action semantics. Journal of Experimental Psychology Human Perception and Performance, 40(5), 1832–1848. 10.1037/a0037449. [DOI] [PMC free article] [PubMed] [Google Scholar]
  87. 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]
  88. Wei T, & Schnur TT (2016). Long-term interference at the semantic level: Evidence from blocked-cyclic picture matching. Journal of Experimental Psychology Learning, Memory, and Cognition, 42(1), 149–157. 10.1037/xlm0000164. [DOI] [PubMed] [Google Scholar]
  89. Wong AL, & Haith AM (2017). Motor planning flexibly optimizes performance under uncertainty about task goals. Nature Communications, 8, 14624 10.1038/ncomms14624. [DOI] [PMC free article] [PubMed] [Google Scholar]
  90. Wong AL, Haith AM, & Krakauer JW (2015). Motor planning.Neuroscientist, 21(4), 385–398. 10.1177/1073858414541484. [DOI] [PubMed] [Google Scholar]
  91. Wurm MF, & Caramazza A (2019). Distinct roles of temporal and frontoparietal cortex in representing actions across vision and language. Nature Communications, 10(1), 289 10.1038/s41467-018-08084-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  92. Wurm MF, & Lingnau A (2015). Decoding actions at different levels of abstraction. Journal of Neuroscience, 35(20), 7727–7735. 10.1523/JNEUROSCI.0188-15.2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  93. Zhang Y, Kimberg DY, Coslett HB, Schwartz MF, &Wang Z (2014). Multivariate lesion-symptom mapping using support vector regression. Human Brain Mapping, 35(12), 5861–5876. 10.1002/hbm.22590. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

1
NIHMS1536109-supplement-1.pdf (2.4MB, pdf)

RESOURCES