Regional changes in word-production laterality after a naming treatment designed to produce a rightward shift in frontal activity

Abstract

Five nonfluent aphasia patients participated in a picture-naming treatment that used an intention manipulation (opening a box and pressing a button on a device in the box with the left hand) to initiate naming trials and was designed to re-lateralize word production mechanisms from the left to the right frontal lobe. To test the underlying assumption regarding re-lateralization, patients participated in fMRI of category-member generation before and after treatment. Generally, the four patients who improved during treatment showed reduced frontal activity from pre- to post-treatment fMRI with increasing concentration of activity in the right posterior frontal lobe (motor/premotor cortex, pars opercularis), demonstrating a significant shift in lateraliity toward the right lateral frontal lobe, as predicted. Three of these four patients showed no left frontal activity by completion of treatment, indicating that right posterior lateral frontal activity supported category-member generation. Patients who improved in treatment showed no difference in lateralization of lateral frontal activity from normal controls pre-treatment, but post-treatment, their lateral frontal activity during category-member generation was significantly more right lateralized than that of controls. Patterns of activity pre- and post-treatment suggested increasing efficiency of cortical processing as a result of treatment in the four patients who improved. The one patient who did not improve during treatment showed a leftward shift in lateral frontal lateralization that was significantly different from the four patients who did improve. Neither medial frontal nor posterior perisylvian re-lateralization from immediately pre- to immediately post-treatment images was a necessary condition for significant treatment gains or shift in lateral frontal lateralization. Of the three patients who improved and in whom posterior perisylvian activity could be measured at post-treatment fMRI, all maintained equal or greater amounts of left-hemisphere perisylvian activity as compared to right. This finding is consistent with reviews suggesting both hemispheres are involved in recovery of language in aphasia patients.

Introduction

It has been clear for more than a century that right-hemisphere mechanisms contribute to language recovery in some patients with aphasia. In particular, right-hemisphere lesion after functional recovery in aphasia can cause deterioration of language (Barlow, 1877, Basso et al., 1989, Gowers, 1887), as can temporary right-hemisphere inactivation during Wada tests (Kinsbourne, 1971). More recent functional imaging studies have provided mixed evidence regarding right-hemisphere participation in language for aphasia patients. However, when the extant literature is carefully reviewed, it indicates that both the right and left-hemispheres contribute to language in aphasia (Crosson et al., 2007a). Some evidence suggests that in nonfluent aphasia, right frontal structures are engaged when their left-hemisphere counterpart is damaged (Blank, Bird, Turkheimer, & Wis, 2003). Hence, in chronic nonfluent aphasia, reorganization of some language production functions to the right hemisphere may be desirable. Yet, many patients fail to demonstrate such reorganization, even with conventional therapy (e.g., Kim, Ko, Parrish, & Kim, 2002).

Thus, a naming treatment was designed to stimulate reorganization of word production to the right lateral frontal lobe. The treatment used a complex left-hand movement to initiate picture-naming attempts. The complex left-hand movement involved opening a box and pressing one of four buttons on a device within the box. Pressing the correct button resulted in presentation of a picture to name on a computer monitor while pressing an incorrect button had no result. The rationale for the treatment was that the complex left-hand movement would activate intention mechanisms in the right medial frontal lobe (e.g., Picard & Strick, 1996), which in turn would activate right lateral frontal structures that could participate in picture-naming. This intention treatment was inspired in part by observations that language performance of some patients with parietal lesions improved with an attention manipulation, i.e., moving stimuli into the ipsilesional hemispace (Coslett, 1999). If an attention manipulation improved language performance in patients with parietal lesions, then an intention manipulation (complex left-hand movement) might improve performance for patients with nonfluent aphasia, which is commonly accompanied by frontal lesions (see Crosson et al., 2007b for the detailed treatment rationale).

The new intention treatment was tested in an efficacy study by administering it and a comparison attention treatment to 34 patients with chronic nonfluent aphasia using a cross-over design (Crosson et al., 2007b). For 11 patients with profound naming impairments (picture-naming <20% correct on 40 items with 12 high, 12 medium, and 16 low-frequency target words; Mean Boston Naming Test = 1.50 of 60 items correct, SD = 2.32), there were no differences between treatments, with roughly half of the patients improving in both treatments. However, in the 23 patients with moderate (picture-naming >70% correct on the 40 items; Mean Boston Naming Test = 33.00 items correct, SD = 8.45) to severe (20 ⩽ picture-naming ⩽ 70% correct on the 40 items; Mean Boston Naming Test = 12.82 items correct, SD = 9.48) naming impairments, this intention treatment produced a faster rate of relearning picture names than did the comparison treatment, and produced a significant treatment response in 89% of patients. Further, treatment effects generalized to untrained items. Yet, these findings, no matter how encouraging, do not demonstrate that the treatment had the intended effect of engaging right lateral frontal mechanisms. Crosson et al. (2005) presented preliminary fMRI data showing that one patient who benefitted from the intention treatment re-lateralized frontal mechanisms to the right hemisphere during category-member generation from pre- to post-intention treatment. The frontal activity of a second patient, who also benefitted from treatment, appeared to be completely re-lateralized to the right hemisphere prior to receiving the intention treatment, when image analysis was time locked solely to spoken response as opposed to stimuli that evoked responses.

The current report expands the fMRI inquiry to five of the patients from our previous efficacy study (Crosson et al., 2007b). These patients received fMRI during category-member generation immediately before and immediately after a 6-week course of the intention treatment. The purpose of this fMRI inquiry was to determine whether the intended changes in the neural substrates of word production occurred as a result of the intention treatment. An improved method of image analysis, combining response- and stimulus-locked approaches (explained below) (Crosson et al., 2007a), was used. Further, because normal older persons may demonstrate greater right-frontal activity than younger persons for word-finding tasks (Wierenga et al., 2008), the current report includes for comparison purposes fMRI findings for a normal age-matched control group on the same word-production task. The main hypothesis was that a positive response to the intention treatment would be accompanied by increased lateralization of lateral frontal activity to the right frontal lobe. Specifically, it was hypothesized that patients who showed significant improvement during treatment would demonstrate increased prominence of right lateral frontal activity during word production, as indicated by shift in fMRI lateral frontal laterality indices rightward from pre- to post-treatment. Also, it was hypothesized that patients who showed significant improvement during treatment would demonstrate greater lateral frontal right-hemisphere lateralization post- (but not pre-) treatment than normal age-, education- and gender-matched controls.

Pre- and post-treatment fMRI data also were used to address three additional questions. (1) The putative mechanism by which right lateral frontal mechanisms were engaged is that the intention manipulation (complex left-hand movement) engages right medial frontal intention mechanisms. In turn, these right-hemisphere mechanisms are thought to activate right lateral frontal structures that, as a result of activation, become engaged in language production. This theoretical proposition raises the following question: Is it necessary for medial frontal structures to demonstrate a rightward shift in lateralization for improvement in treatment or for a rightward shift in lateral frontal activity to occur? (2) Data suggest that both right- and left-hemisphere mechanisms may contribute to language recovery (Crosson, 2007, Crosson et al., 2007a). Further, re-lateralization of activity to the right hemisphere often occurs in structures homologous to the damaged structures (e.g., Blank et al., 2003, Calvert et al., 2000, Lazar et al., 2000, Thulborn et al., 1999, Weiller et al., 1995). Such data suggest that even when right-hemisphere structures are engaged in the service of language functions in aphasia, residual knowledge of language in the left-hemisphere is useful for leveraging recovery or treatment gains. This analysis indicates that even if the intention treatment is successful in increasing right-frontal participation in language production, it may also be necessary to use residual language knowledge in the left posterior perisylvian cortex, raising the following question: Does left posterior perisylvian cortex remain active after the intention treatment induces a rightward shift in lateral frontal activity during word production? (3) Some rehabilitation studies suggest that activity may become more focused and less diffuse as rehabilitation proceeds (e.g., Dobkin, Firestine, West, Saremi, & Woods, 2004). Indeed, repetitive transcranial magnetic stimulation (rTMS) data suggest that right pars triangularis activity actually may impede language recovery, even though activity in more posterior right inferior frontal cortex (right pars opercularis) may support language recovery (Naeser et al., 2005). These data raise the following question: Will post-intention treatment activity become more focused than pre-treatment activity and will activity in right frontal cortex be confined to structures posterior to pars triangularis?