Spatial neglect treatment: the brain’s spatial-motor Aiming systems

Abstract

Animal and human literature supports spatial-motor “Aiming” bias, a frontal-subcortical syndrome, as a core deficit in spatial neglect. However, spatial neglect treatment studies rarely assess Aiming errors. Two knowledge gaps result: spatial neglect rehabilitation studies fail to capture the impact on motor-exploratory aspects of functional disability. Also, across spatial neglect treatment studies, discrepant treatment effects may also result from sampling different proportions of patients with Aiming bias. We review behavioral evidence for Aiming spatial neglect, and demonstrate the importance of measuring and targeting Aiming bias for treatment, by reviewing literature on Aiming spatial neglect and prism adaptation treatment, and presenting new preliminary data on bromocriptine treatment. Finally, we review neuroanatomical and network disruption that may give rise to Aiming spatial neglect. Because Aiming spatial neglect predicts prism adaptation treatment response, assessment may broaden the ability of rehabilitation research to capture functionally-relevant disability. Frontal brain lesions predict both the presence of Aiming spatial neglect, and a robust response to some spatial neglect interventions. Research is needed that co-stratifies spatial neglect patients by lesion location and Aiming spatial neglect, to personalize spatial neglect rehabilitation and perhaps even open a path to spatial retraining as a means of promoting better mobility after stroke.

Introduction.

In the first weeks and months after a stroke, over 50% of patients demonstrate pathologically asymmetric spatial behavior causing functional disability, defined as spatial neglect (Barrett & Burkholder, 2006). With an ever-increasing number of stroke survivors requiring post-acute care (Feigin et al., 2017), the need for effective spatial neglect rehabilitation is growing. In this review, we argue for the importance of targeted treatment in spatial neglect. Specifically, we 1) review behavioral evidence for spatial-motor Aiming bias as a distinct set of spatial neglect symptoms in both human and non-human animals; 2) demonstrate the importance of measuring Aiming spatial neglect and targeting its treatment; and 3) review neuroanatomical and network disruption that may give rise to Aiming spatial neglect, including evidence from non-invasive brain stimulation and neurophysiology.

Biological models of spatial neglect in both humans and animals suggest the presence of distinct spatial cognitive information processing stages. Any one or any combination of these stages may be disrupted in an individual presenting with spatial neglect (Heilman et al., 2003; Mesulam, 2000; Rengachary, He, Shulman, & Corbetta, 2011; Verdon et al., 2010). Behaviorally, perceptual-attentional, input-associated “Where” spatial cognitive processing can be dissociated from motor-intentional, output-association “Aiming” spatial cognitive processing (Barrett, 2014; Bisiach, Geminiani, Berti, & Rusconi, 1990). Impairment at relatively early, input-related stages of information processing may lead to deficits in the spatial distribution of attention, affecting stimulus encoding. These spatial cognitive input abnormalities are defined as “Where” spatial neglect. “Where” deficits include a reduced ability to allocate perceptual resources across the entire spatial field, difficulty in focusing spatial attention, or difficulty in attentional disengagement (Barrett, Beversdorf, Crucian, & Heilman, 1998; Barrett, Schwartz, Crucian, Kim, & Heilman, 2000; Rapcsak, Verfaellie, Fleet, & Heilman, 1989).

Aiming spatial neglect is defined by impairment at relatively late, output-related stages of information processing, which may disturb the representation of spatial information supporting action, or, more commonly, may directly affect both action planning and action execution (Heilman, 2004; Heilman et al., 2003; Hillis, 2006). These spatial cognitive output abnormalities are defined as “Aiming” spatial neglect. Aiming deficits include a slowness or failure to initiate movements in or toward contralesional space (e.g., Heilman et al. 1985) and may also include a reduction in the magnitude of contralesionally-directed motor activities (e.g., Bisiach et al, 1990). They can affect movements of the limbs, body, and eyes (De Renzi, Colombo, Faglioni, & Gibertoni, 1982; Ringman, Saver, Woolson, & Adams, 2005), resulting in pathologically asymmetric movement, action and intention causing functional disability (Riestra & Barrett, 2013). Among the most prominent symptoms of Aiming spatial neglect is directional hypokinesia, defined as a decreased propensity or disinclination to move any part of the body in one spatial direction. In addition, Aiming spatial neglect may manifest in a family of symptoms described by terms that include “motor-intentional” (Heilman, 2004; Luaute et al., 2018) or “premotor” neglect (Bisiach et al., 1990). Distinct from, but closely related to directional hypokinesia, is hemispatial hypokinesia (movements that are asymmetric in force, gain or frequency when performed in the neglected body space, versus the non-neglected body space). Lastly, limb hypokinesia, or failure to activate the limbs on the neglected side of the body, with respect to the frequency, force, or gain of movements, is another Aiming spatial neglect manifestation (Laplane and Degos, 1983). The interested reader is encouraged to consult reviews describing Aiming spatial neglect and its associated deficits (Barrett, 2014; Barrett & Houston, 2019; Champod, 2020; Chaudhari, Pigott, & Barrett, 2015).

Finally, spatial cognition also entails an intermediate information processing stage, in which explicit and implicit spatial representations, knowledge in the form of imagery, maps and internal descriptions, are required. These can be visual-spatial, auditory-spatial, and even spatial somesthetic representations, any of which may be impaired in spatial neglect.

At present, guideline-based spatial neglect rehabilitation ignores the potential heterogeneity in neglect presentation among post-stroke individuals. “Where” spatial neglect and its perceptual-attentional errors are often the only treatment target (Barrett, 2014; Barrett & Burkholder, 2006). Studying changes in spatial Aiming errors is likely to serve as an excellent route to collecting data on spatial neglect treatment that has a valid relationship to clinical status and daily life function (Barrett, Goedert, & Basso, 2012). Further, developing a concept of Aiming spatial neglect may provide a basis for specifically effective treatments to enhance safety, activity and function (Barrett & Muzaffar, 2014) in the first weeks and months of stroke recovery. Such treatments would represent a unique and personalized approach.

In the first part of this paper, we review previous studies, which demonstrated that Aiming spatial neglect characteristics predict functional recovery of spatial neglect in response to treatment. Dopaminergic medications have also been proposed to improve spatial neglect. In the second part of this paper, we will present data from a preliminary study. This data suggests that, in patients with functionally-relevant spatial neglect symptoms on performance testing, a laboratory measure of Aiming spatial neglect predicts functional improvement in response to bromocriptine. As we demonstrate in this preliminary study, selective effects of a treatment on Aiming spatial neglect can be detected by utilizing mechanistic (brain system-relevant) measures of spatial neglect as predictors, with functional (disability-relevant) outcomes, examined pre- and post-treatment.

Behavioral Evidence for a Distinct Spatial-motor Aiming System.

Over almost 50 years, strong evidence of spatial-motor Aiming bias as part of spatial neglect has been available from both human and animal studies (Barrett, 2014) (Heilman, 2004).

Animal models.

Across the mammalian class, spatial neglect, and specifically Aiming spatial neglect, can be produced by ablating subcortical ascending dopaminergic input to the brain, or frontal lobe motor regions (Apicella, Trouche, Nieoullon, Legallet, & Dusticier, 1990; Deuel, 1992; Schwarting & Huston, 1996). In these studies, animals manifest spatial Aiming bias by spontaneously rotating in the ipsilesional direction (Urban Ungerstedt, 1976), a form of directional hypokinesia. Ungerstedt and colleagues (Urban Ungerstedt, 1976; U. Ungerstedt & Arbuthnott, 1970; U. Ungerstedt, Ljungberg, & Steg, 1974; Zetterstrom, Herrera-Marschitz, & Ungerstedt, 1986) made the key observation that midbrain dopaminergic depletion was associated with “sensory neglect” and directional hypokinesia in rats, in the absence of hemiparesis. This research links spatial neglect to both motor networks and subcortical neuroanatomic processing.

This spontaneous rotation is not consistent with a deficit of awareness, perception or attention. Hoyman and colleagues (Hoyman, Weese, & Frommer, 1979) trained rats to respond by orienting to the side opposite a tactile stimulus, and then unilaterally lesioned the nigrostriatal system. The rats failed to orient contralesionally to tactile stimulation of the ipsilesional, good side; orienting ipsilesionally to contralesional stimulation was not impaired. This indicates that the spatial neglect was attributable to a failure of contralesional spatial movement—a spatial-motor Aiming deficit. Carli et al. (Carli, Evenden, & Robbins, 1985) later replicated this finding after lesioning the rat caudate, demonstrating impaired initiation of contralateral responses, but not detection of contralateral stimuli. Monkeys with unilateral toxic ablation of the substantia nigra from carotid administration of 1-methyl-4-phenyl 1,2,3,6-tetrahydropyridin demonstrate slower movements in the contralesional than ipsilesional direction with the ipsilesional hand and arm, and slowed movements of reduced frequency with the contralesional limb as compared with the ipsilesional limb (Schneider, McLaughlin, & Roeltgen, 1992). In another experiment, after toxic unilateral dopaminergic ablation, resulting in 90% loss of cells in the caudate, putamen, nucleus accumbens, and substantia nigra, monkeys showed more body rotations toward the lesion side and persistent head deviation toward the ipsilesional side. In addition, these animals had problems reaching into contralesional space, using both the good and the impaired arm (Milton, Marshall, Cummings, Baker, & Ridley, 2004).

Thus, although involvement of motor networks in spatial neglect does not automatically mean that the dopaminergic system is impaired in humans, there is ample evidence that dopaminergic disruption induces Aiming spatial neglect symptoms in animals. We also see the converse pattern of evidence: The effect of pharmacologic modulation on directional hypokinesia and Aiming spatial neglect is also clearly demonstrated in animals (Fuxe & Ungerstedt, 1976; Herrera-Marschitz & Ungerstedt, 1984, 1985; T. E. Robinson & Becker, 1983; Schultz & Ungerstedt, 1978), with dopaminergic and noradrenergic stimulation reducing pathologic rotational behavior, or even reversing its pattern.

Human Aiming spatial neglect.

As reviewed above, symptoms Aiming spatial neglect includes a group of related symptoms, including directional hypokinesia, hemispatial hypokinesia, and limb hypokinesia (Heilman, 2004; Hillis et al., 2006; Laplane & Degos, 1983). Directional hypokinesia is a common symptom of Aiming spatial neglect in humans, analogous to spontaneous rotation or circling behavior in animals. With directional hypokinesia there is a propensity to move much more readily in one spatial direction, compared to the opposite direction. When weakness is present, movement asymmetry in directional hypokinesia must exceed the asymmetry induced by mechanical forces to be considered a spatial deficit. Patients with directional hypokinesia after right brain stroke generally demonstrate leftward hypokinesia: they readily move rightward, but have difficulty moving leftward. Directional hypokinesia can affect either the right or left hand, the head, eyes, or trunk and whole-body movements (Barrett & Muzaffar, 2014; Bisiach et al., 1990; Na et al., 1998)(Heilman et al., 1985). In addition to reduced amplitude of leftward movement after right stroke, patients may generate leftward movements that are slowed relative to rightward movements, or of reduced force compared to rightward movement. Aiming spatial neglect can affect the eyes, either limb, or whole-body movements after stroke (Adair & Barrett, 2008; Barrett, 2017; Heilman, 2004). Although hemispatial hypokinesia and limb hypokinesia are distinct symptoms of Aiming spatial neglect, they can occur separately or together. Their close relationship is supported by co-manipulation by motor-intentional spatial neglect interventions (Fortis, Chen, Goedert, & Barrett, 2011; Hreha, Gillen, Noce, & Nilsen, 2018)

A barrier to understanding the impact of Aiming spatial neglect and motor-exploratory asymmetry in humans is that these symptoms are usually not measured or diagnosed during clinical care. Standard paper and pencil assessment for attentional disorders include the Behavioral Inattention Test-conventional subtest, which may preferentially assess input-related, Where, deficits in spatial attention (Goedert et al., 2012). Validated assessment of actual adaptive movements during functional performance with measures like the Catherine Bergego Scale (Azouvi et al., 2003; Chen, Chen, Hreha, Goedert, & Barrett, 2015; Mark, 2003), is much more likely to detect disruption to Aiming spatial-motor systems.

Nonetheless, these standard clinical assessments cannot perfectly discriminate between input- and output-components of spatial cognition. In contrast, laboratory measures can measure Aiming bias with high reliability and consistency in healthy subjects (Chen, Erdahl, & Barrett, 2009; Chen et al., 2011; Garza, Eslinger, & Barrett, 2008; Goedert, Chen, Foundas, & Barrett, 2020; Schwartz, Adair, Na, Williamson, & Heilman, 1997; Shah, Gonzalez, & Barrett, 2012) and can objectively quantify directional hypokinesia (Barrett & Burkholder, 2006; Fortis et al., 2011; Goedert, Chen, Boston, Foundas, & Barrett, 2014; Goedert et al., 2012; Goedert et al., 2020; Khurshid, Longin, Crucian, & Barrett, 2009). In this laboratory measure participants used a mouse to bisect horizontal lines appearing on a computer screen. On half the trials, we reversed the feedback of the mouse movement, such that when the subject moved the mouse leftward, it appeared to move rightward and vice versa. This procedure allows us to dissociate motor-intentional Aiming errors from perceptual-attentional Where errors. A Where spatial deficit is characterized by errors that change direction with video feedback reversal. In contrast, an Aiming spatial bias is characterized by errors that do not change direction. The precise quantity of Where and Aiming spatial error contribution to total line bisection error can be separately quantified. Furthermore, we demonstrated that this measure of Aiming spatial neglect has construct validity, by observing that this kind of bias changes specifically and systematically with dopaminergic input, motor cuing and motor adaptation (Barrett, Crucian, Schwartz, & Heilman, 1999; Fortis et al., 2011; Garza et al., 2008; Schwartz, Barrett, Kim, & Heilman, 1999). Where bias, in contrast, is specifically altered by visual distraction and sensory deprivation (Barrett & Burkholder, 2006; Goedert et al., 2012; Khurshid et al., 2009) (Sacchetti, Goedert, Foundas, & Barrett, 2015)

Researchers who used this assessment, or similar laboratory-based measures of Aiming bias, demonstrated that human Aiming errors may be depend on physiological and behavioral factors that are similar to the factors that alter rotational behavior in animals (Barrett et al., 1999; Fleet, Valenstein, Watson, & Heilman, 1987; Geminiani, Bottini, & Sterzi, 1998; Gorgoraptis et al., 2012). Even in healthy people, differences in the direction and magnitude of spatial orienting bias measured using a grayscales task were strongly associated with the pattern of asymmetric left/right binding of dopamine (DA) D2 receptors in the striatum, frontal and temporal cortex. This asymmetry in tonic DA may confer a hemispheric advantage reflected in orienting bias (Tomer et al., 2013). Motor-related interventions also result in changes to other forms of Aiming spatial neglect in humans. Geminiani (Geminiani et al., 1998) observed that although verbal report of visual stimuli did not improve after dopaminergic treatment with apomorphine, ballistic pointing to left-sided stimuli specifically improved, which likely reflected improvement in left hemispatial hypokinesia (Heilman, 2004). A third manifestation of Aiming spatial neglect, limb hypokinesia (measured by force generated by the contralesional hand) responded in a prior study to a spatial neglect intervention (optokinetic stimulation) (Vallar, Guariglia, Nico, & Pizzamiglio, 1997).

Functional relevance of measuring Aiming spatial neglect.

Aiming spatial neglect predicts functional recovery in response to spatial neglect treatment.

Both functional (disability-relevant) and mechanistic effects of spatial neglect rehabilitation may be specific to Aiming spatial neglect. We (Goedert et al., 2012) identified a subset of items on the functional performance assessment of spatial neglect, the Catherine Bergego Scale, which potentially identified motor-exploratory bias. These items independently accounted for an additional 9% of the variability in functional disability measured with the Barthel Index, beyond that accounted for by traditional paper and pencil assessment.

Importantly, while guideline-based care pathways recommend approaches for spatial neglect treatment such as visual scanning training, limb activation, and prism adaptation treatment, these guidelines treat neglect as a unitary entity. However, evidence from our laboratories and those of others suggest that some of these treatments may specifically target different stages of information processing in spatial cognition. In particular, in our laboratory we have demonstrated that prism adaptation appears to exert a direct effect on spatial Aiming bias (specifically, on directional hypokinesia), as compared with spatial Where unawareness (Fortis, Chen et al., 2011). In some experiments, spatial neglect interventions, including prism adaptation, improved spatial motor Aiming bias but actually resulted in increased spatial Where, perceptual-attentional error. (Barrett & Burkholder, 2006; Fortis et al., 2011; Schwartz et al., 1999). In other words, spatial neglect interventions had opposing effects on Aiming spatial neglect and Where spatial neglect symptoms. This means that unless we include measures sensitive and specific to changes in Aiming spatial neglect, global outcome measures in spatial neglect rehabilitation are likely to reflect a combination of improvement of one type of bias, and worsening of another type of bias, and may thus be contaminated and invalid. If we do not assess Aiming spatial neglect, we also may fail to diagnose neglect-related functional disability that solely affects spatial Aiming, or we may also observe, but misinterpret, these symptoms (Barrett, 2017).

Another way of viewing the Aiming spatial neglect diagnosis gap is that it may explain the paradox of persistent disability in “recovered” spatial neglect. Although spatial neglect as assessed on paper and pencil tests may resolve completely, disability frequently persists in patients who had spatial neglect acutely: for example, reduced community mobility (Oh-Park et al., 2014). Measuring Aiming spatial neglect may help to explain these chronic limitations to activity and participation, since “recovery” on paper and pencil assessment may be highly reflective of changes in Where spatial function, without reflecting the status of spatial motor-exploratory and Aiming spatial function (Goedert et al., 2012).

A guideline-based treatment approach to spatial neglect, versus Aiming spatial neglect treatment.

We recently wrote about the differences between utilizing a personalized medicine approach to treating spatial neglect, based on behavioral spatial neglect symptoms, and utilizing guideline-based care pathways, which treat spatial neglect as a unitary entity (Barrett, Abdou, & Caulfield, 2019). Excellent practice guidelines and discipline-specific systematic review documents are available (Gillen et al., 2015; Intercollegiate Stroke Working Party, 2016; Winstein et al., 2016), and these are very helpful to list options for clinicians in evidence-based spatial neglect interventions. Among these approaches are visual scanning training, limb activation, and prism adaptation treatment (Barrett et al., 2012; Chen, Pitteri, Gillen, & Ayyala, 2017; Eskes, Butler, McDonald, Harrison, & Phillips, 2003; Robertson, Hogg, & McMillan, 1998; Weinberg et al., 1977). However, as we wrote (Barrett et al., 2019), page 140), “‘evidence-based’ approaches recommended for spatial neglect patients did not take into account the differences in response to treatment that might be observed across patients with different symptom patterns … This is an overall problem with using randomized controlled trials, meta-analyses, and systematic reviews to determine the care of individual patients: a treatment with best outcomes for groups of average patients is often a poor choice for a patient who is atypical in one or more major characteristics (Berguer, 2004).”

Preliminary Data suggests Aiming spatial neglect predicts functional recovery of spatial neglect in response to bromocriptine treatment.

Rationale.

Animal and human studies suggest that spatial Aiming bias and spatial neglect may be altered after administering dopaminergic medications ((Geminiani et al., 1998; U. Ungerstedt & Arbuthnott, 1970). In this section of the paper, we present preliminary data we collected in collaboration with Priyanka Shah-Basak to evaluate the following objective. When stroke survivors with spatial neglect were administered dopaminergic medication, we predicted we would observe improvement in spatial neglect-specific functional performance when survivors had Aiming spatial neglect, which would exceed those observed in those survivors with Where spatial neglect. In this preliminary study, we observed that patients with Aiming spatial neglect responded favorably to bromocriptine, with sustained improvement in functional disability. However, patients with Where deficits, and those who had both Where and Aiming spatial deficits, did not improve in spatial neglect-related functional disability over the 7-week period (Figure 1).

Figure 1:

CBS scores for individual recovery trajectories (gray lines) and group-averaged recovery trajectory (black line) for (A) participants with Aiming bias, (B) Where bias, and (C) both biases. On the CBS, lower scores represent less impairment.

Subjects.

Demographic information.

Based on recruitment feasibility, a sample of ten right hemispheric stroke survivors (5 women, mean age = 60.2 ± 11.06 years, 5 to 14 days post-stroke, mean education 13.4 ± 4.8). Subjects were clinically screened for spatial neglect after referral of their admitting physician or therapy team during inpatient rehabilitation, based on clinical suspicion of this diagnosis (2015–2016). The diagnosis of spatial neglect was confirmed based on a score ≥ 2 on functional performance assessment with the Catherine Bergego Scale (CBS;(Chen, Hreha, Fortis, Goedert, & Barrett, 2012). We chose this diagnostic criterion because of its predictive validity to diagnose functional disability in spatial neglect, and not simply the presence or absence of spatial impairment in an office or laboratory setting. Participants gave written informed consent for this study, as required by the local Institutional Review Board. All subjects enrolled were retained to complete all assessments in the study.

Materials and methods.

Aiming and Where spatial neglect were assessed at baseline upon study entry using the computerized line bisection assessment with video feedback reversal, as described in section 2a (Chen et al., 2009). Unlike the paper-and-pencil line bisection task, which measures errors that are induced by a combination of both Where and Aiming spatial bias, and does not allow these components to be fractionated, this controlled, mechanistic task has been convergently validated in healthy subjects, via its relationship to top-down motor cuing with ballistic pointing (Garza et al., 2008). The paper and pencil line bisection lacks sensitivity when diagnosing spatial neglect (Ferber & Karnath, 2001); however, the laboratory test was used for behavioral subtype characterization; the Catherine Bergego Scale, as above, was used for diagnosis.

Bromocriptine, a D2 dopamine receptor agonist, was orally administered twice daily to all participants for seven weeks. Dosage was increased based on a pre-specified study protocol and prior studies of bromocriptine for spatial neglect (Barrett et al., 1999; Fleet et al., 1987), from 1.25 mg daily to 7.5 mg twice daily within the first 3 weeks, and maintained at 15 mg/day for four more weeks. We assessed the severity of spatial neglect using the Catherine Bergego Scale, (CBS)(Chen et al., 2012) weekly for each of the seven weeks. This scale has demonstrated validity to predict functional disability (Azouvi, 2016).

Objective:

To test the hypothesis that bromocriptine would have a different effect on participants with functional, disability-relevant Aiming spatial neglect, than it did in those with baseline Where spatial neglect, we evaluated whether the trajectory of recovery of spatial neglect on the Catherine Bergego Scale was different in stroke survivors who made Aiming spatial errors at baseline, versus those who made only Where spatial errors.

Statistical analysis:

Using results from the computerized line bisection task, and using a categorization method from a previous stratification study (Goedert et al., 2014), participants were split into 3 groups – those with only Where spatial deficits (n=3, median Baseline CBS = 23), those with only Aiming spatial deficits (n=4, median baseline CBS = 20), and those with both Where and Aiming spatial deficits (n=3, median baseline CBS = 23). To test our a priori hypothesis, we performed a multi-level modeling (MLM) analysis (Goedert, Boston, & Barrett, 2013), controlling for baseline CBS scores, to assess for improvements in spatial neglect, as assessed by the CBS, over the seven-week bromocriptine administration period.

Results:

The MLM analysis revealed that analysis of spatial Aiming and Where bias in baseline functional performance predicted functional improvement. In particular, there was an interaction between baseline bias and week, F(2, 43) = 4.79, p = 0.013. For this model, the marginal R², which represents solely the contributions of the fixed effects, was 0.43. Compared to participants with Where spatial deficits, those with Aiming spatial deficits had the steepest linear recovery (decline in functional disability) on the CBS (b=−0.21, SE=0.08, p=0.008), over 7 weeks (See Figure 1). Participants with both types of spatial deficits (Where + Aiming) had a similar linear recovery slope to those with Where deficits alone (b=−0.001, SE=0.10, p>0.05), and both the Where and the Where + Aiming groups had a linear recovery slope not significantly different from zero (ps > .05).

Although the study subgroups in this preliminary study were small, the results suggest that, consistent with patterns observed with prism adaptation treatment (Goedert et al., 2014), bromocriptine treatment may selectively improve Aiming spatial neglect, leaving Where spatial neglect unaffected. The results are suggestive that administering bromocriptine to patients selected based on their spatial neglect subtype (biomarker-targeted treatment administration) can be developed based on these results. This difference in response between patients with Aiming spatial neglect and those with both Where and Aiming bias may explain negative treatment findings in prior studies of dopaminergic therapies for spatial neglect (e.g. Buxbaum et al., 2007). In previous group-averaged studies, benefit to patients with Aiming spatial deficits may have been canceled out by adverse or null effects for patients who had Where spatial deficits.

We strongly emphasize that these results should be viewed as preliminary, because the small size of the study sample limits generalizability to other patients with spatial neglect. While this initial, preliminary study seeking specificity of a bromocriptine treatment signal to spatial Aiming bias did not specifically adjust for other confounders such as spontaneous recovery, age, depression, or other medications, future large-scale studies could address these gaps. It may be that Aiming spatial neglect was predictive of reduced functional performance problems with bromocriptine therapy because of the presence of frontal brain lesions, as has been suggested by prior studies of prism adaptation treatment (Goedert et al., 2020). Thus, future large-scale studies should explicitly examine the three-factor relationship (Aiming spatial neglect, frontal lesion location, and recovery with bromocriptine). A future, large-scale clinical trial could enroll only the subgroup of patients with Aiming spatial neglect, and randomize these patients to receive either bromocriptine or no medication, stratifying patients by lesion location. This can be considered, now that this preliminary study suggests strong feasibility of a treatment study targeted to patients with Aiming spatial neglect. In future studies, it may be interesting to examine eye movement changes, and changes in the brain network supporting eye movements, because these systems may also be influenced by Aiming spatial neglect (Barrett, 2017). Future studies should also evaluate awareness of deficit as a study outcome, which is clearly important to functional disability when patients have Aiming spatial neglect, and can improve with interventions (Cappa, Sterzi, Vallar, & Bisiach, 1987).

Bromocriptine may specifically target spatial Aiming, motor-intentional networks associated with directional action, response inhibition, persistence, and motor/personal self- regulation. It is also possible that bromocriptine may more selectively benefit patients with mild neglect; participants with predominant Aiming deficits had milder neglect symptoms in our study.

Network Models of Spatial Aiming in Neglect

While results from traditional lesion-symptom mapping studies emphasize the explanatory power of local neuroanatomy, in examining the neurobiological basis of spatial Aiming bias, we suggest that investigators strongly consider the distributed injury hypothesis (Corbetta, Kincade, Lewis, Snyder, & Sapir, 2005). This hypothesis generally states that focal lesions can have distant effects within a distributed brain network by virtue of diaschisis, altering functional connectivity, damaging long-range white matter tracts, or disrupting neurotransmitter systems with broad projections. While the network view of brain function makes it much more challenging to understand the effects of lesions on behavior, it may also offer more avenues for intervention and rehabilitation. Unfortunately, existing neurophysiologic network models are not yet available to fully explain how derangement of network interaction would give rise to the Aiming spatial neglect described above. In this section we review extant evidence for how spatial motor aiming dysfunction may arise in the mammalian brain.

Noninvasive brain stimulation.

Regional specialization within the human fronto-parietal network is not yet well understood (Filimon, 2010) (Archambault, Ferrari-Toniolo, Caminiti, & Battaglia-Mayer, 2015) (Battaglia-Mayer, Babicola, & Satta, 2016), but transient disruption of regional cortical activity using repetitive transcranial magnetic stimulation in healthy subjects may provide some insight into the relative contributions of motor and parietal cortex to Aiming spatial neglect (Chouinard & Paus, 2010).

In a review of fifty studies where repetitive transcranial magnetic stimulation (rTMS) was delivered to premotor cortical regions, the cumulative results suggested that the dorsal (PMd) and ventral (PMv) premotor cortex, the supplementary motor area (SMA) and pre-supplementary motor area (pre-SMA) are primarily involved in response selection, grasping, movement sequencing/bimanual coordination and response inhibition/task switching, respectively (Chouinard & Paus, 2010). In studies investigating motor biases, right frontal and right parietal rTMS both have been used to induce symptoms of left neglect in a line bisection task, but only frontal inhibitory TMS caused a spatial Aiming bias specifically (Ghacibeh, Shenker, Winter, Triggs, & Heilman, 2007). In a more recent study of the long-lasting offline effects of repetitive TMS, inhibition of the right posterior middle frontal gyrus (pMFG) reduced exploratory visuo-motor behavior as measured with target cancellation tests (Platz, Schuttauf, Aschenbach, Mengdehl, & Lotze, 2016). Also, rTMS to the right middle frontal gyrus (rMFG) in healthy individuals using their right hand without visual feedback was found to reduce the number of internally generated movements directed toward the left side. However, reaction times, speed of movement and pointing accuracy remained unaffected. The authors concluded that rMFG disruption led to a decreased likelihood of completing leftward movements (Gutierrez-Herrera, Saevarsson, Huber, Hermsdorfer, & Stadler, 2017). However, another study reported that only left PMd (not right PMd) TMS led to an abnormal overshoot of right hand reaches (Davare, Zenon, Desmurget, & Olivier, 2015). Thus, TMS studies in humans make a compelling argument for spatial motor Aiming biases being associated with frontal lobe lesions, possibly involving PMd in particular.

Primate neurophysiology of Aiming spatial neglect.

In primates, activation of different parts of PMd, PMv, and SMA during reaching have been extensively studied (Hoshi & Tanji, 2002; Picard & Strick, 2003; Romo & Schultz, 1992; Suminski, Mardoum, Lillicrap, & Hatsopoulos, 2015; Takahashi, Best, & Huh, 2017; Xiao, Padoa-Schioppa, & Bizzi, 2006). Intracortical microstimulation (ICMS) of the premotor cortex has been shown to elicit not only muscle twitches in the upper extremity but also more complex movements (Graziano, Taylor, & Moore, 2002). These regions may even instruct motor learning (Mazurek & Schieber, 2017).

Frontal-subcortical networks.

Because spatial neglect has traditionally been considered a cortical syndrome, it is important to consider its relationship to frontal lobe deficits. As above, frontal cortical inhibition has been linked to Aiming bias, however a significant amount of experimental data also supports the association between directional motoric deficits and frontal lobe lesions. It has been appreciated for nearly five decades that frontal lobe lesions can lead to Aiming spatial neglect (Castaigne, Laplane, & Degos, 1972; Watson, Miller, & Heilman, 1978). In one study, reaching accuracy was decreased by regional inactivation with the injection of the GABA agonist muscimol into the PMd, but it was reaching speed that was decreased when the PMv was injected (Kurata & Hoffman, 1994). In another study, action initiation was affected by injection of the GABA antagonist bicuculline into the PMd, causing animals to reach before the expected cue (Sawaguchi, Yamane, & Kubota, 1996). Action selection may also be affected by premotor cortex activity. For instance, when a monkey that underwent unilateral PMv inactivation using muscimol was presented with two equivalent food morsels simultaneously to the monkey’s right and left, the monkey was less likely to reach for the morsel contralateral to the side of inactivation (Schieber, 2000). Overall, the animal’s behavior had many of the features characteristic of directional hypokinesia. This particular paradigm may be an important model of behaviors in humans with right frontal lobe damage and Aiming SN as they are interacting with their environment. Although the above ablation studies are usually interpreted as indicating that frontal cortical regions support healthy spatial Aiming, these studies could also suggest that disconnection of the frontal cortex from its associated subcortical or posterior regions could be linked to spatial movement asymmetry, and thus may contribute to Aiming spatial neglect.

Aiming Spatial neglect and brain networks.

Translational evidence.

In combined experimental data in humans and in animals, spatial motor Aiming bias may be associated with lesions in the frontal lobe, also with subcortical lesions of the caudate, putamen or superior colliculus, and much less commonly with parietal lesions. In one study of 15 patients with right brain damage mostly due to ischemic stroke, subjects with the highest scores on a measure of motor Aiming bias had mostly pre-rolandic frontal lobe lesions; two others had subcortical lesions (Bisiach et al., 1990). This general finding has since been replicated in multiple studies (Coslett, Bowers, Fitzpatrick, Haws, & Heilman, 1990; Na et al., 1998; Tegner & Levander, 1991). Sapir et al (2007) sought to identify the neural correlates of directional hypokinesia in right hemisphere stroke patients with spatial neglect by comparing neglect patients with and without motor bias (Sapir, Kaplan, He, & Corbetta, 2007). They found that damage of the white matter underlying the frontal lobe precentral gyrus, inferior frontal gyrus, frontal operculum, and anterior insula was associated with motor bias. This lesion distribution is consistent with that found in studies of neglect patients with Aiming bias who improve in response to treatment with prism adaptation therapy (Chen, Goedert, Shah, Foundas, & Barrett, 2014; Fortis et al., 2011; Goedert et al., 2014; Goedert et al., 2020).

Training with 20-diopter, yoked wedge optical prisms that shift the viewed visual field rightward improves functional abilities in spatial neglect (see systematic review in Champod et al., 2016). This protocol of prism adaptation also alters spatial-motor Aiming bias in healthy controls and people with spatial neglect (Fortis et al., 2011a; Fortis et al. 2011b). When our research group noted altered spatial Aiming bias with prism adaptation therapy, this led us to stratify patients so that we made a further discovery: greater improvements in spatial functional performance occurs after prism adaptation therapy in patients with spatial motor Aiming bias and frontal cortical lesions (Goedert et al., 2014; Chen et al., 2014; Goedert et al., 2020).

Subcortical neuroanatomic systems and Aiming spatial neglect.

Damage to subcortical structures has also been associated with spatial Aiming bias. In the same study of the neural correlates of directional hypokinesia that implicated frontal lobe regions, Sapir et al (2007) also reported a strong association of this bias with damage of the ventral lateral putamen and claustrum, and speculated that this might result in a disruption of subcortical dopaminergic neurotransmission which can play an important role in spatial cognition as discussed above (Sapir et al., 2007). Another subcortical structure, the mammalian superior colliculus (SC) is a major hub of sensorimotor integration and neurons in deep layers of the SC are crucial to the generation of contraversive saccadic eye movements (D. A. Robinson, 1972). In addition, the SC has other motoric functions (Gandhi & Katnani, 2011), including multiple classes of neurons that are active prior to and during arm movements, and which may contribute to the control of visually-guided movements (Lunenburger, Kleiser, Stuphorn, Miller, & Hoffmann, 2001). Neuroimaging studies also provided evidence for reach-related neurons (Himmelbach, Linzenbold, & Ilg, 2013; Linzenbold & Himmelbach, 2012). Persons with spatial neglect make fewer saccades into the neglected field. When the neurotoxin MPTP was infused into the head-body junction of the caudate nucleus of monkeys, spontaneous saccades became less frequent, with decreases in amplitudes and peak velocities that were more apparent for saccades directed toward the contralesional side (Kato et al., 1995). However, more research needs to be done to understand directional hypokinesia of eye movements in humans with spatial neglect, since some authors maintained saccade speeds and amplitude can be normal in spatial neglect (Behrmann, Ghiselli-Crippa, & Dimatteo, 2001) and that motoric bias against contralesional saccades can be due to impaired initiation. It is not clear whether patients with impaired initiation and normal contralesional saccades studied in that report might have other types of spatial Aiming deficits, however, such as limb hypokinesia or hemispatial hypokinesia.

In spite of evidence implicating the premotor cortex, there is yet no final consensus on the critical neuroanatomy and fundamental mechanism driving spatial motor Aiming bias (Saevarsson, Eger, & Gutierrez-Herrera, 2014). In Aiming spatial neglect, prefrontal cortex function will likely need to be understood within the context of its network of connectivity with its parietal counterpart; they are inextricably linked via multiple large white matter fiber tracts like the superior longitudinal fasciculus (SLF). Within a distributed network, a focal lesion may have wide-spread repercussions as the result of diaschisis-like phenomena (Carrera & Tononi, 2014), changes in functional connectivity (Lim & Kang, 2015), and damage to long-range white matter tracts like the SLF (Doricchi, Thiebaut de Schotten, Tomaiuolo, & Bartolomeo, 2008) (Thiebaut de Schotten et al., 2014), or the disruption of the balance of cholinergic and dopaminergic neurotransmission (Luvizutto et al., 2015). Also, for lesion-symptom mapping studies to better inform mechanistic models, errors on behavioral tests that are due to biased movement need to be distinguishable from errors due to unilateral unawareness; this can be a problem when the workplace for stimuli and the workplace for movements are fully aligned. Finally, a true understanding of directional motor biases will require that we are able to consider several neurophysiological fundamental sources of asymmetry. Posterior parietal cortex is a) involved in motor planning and b) may support improved movement after rehabilitation in patients with spatial neglect (Mattingley, Husain, Rorden, Kennard, & Driver, 1998). We also need to measure electrophysiologic, hemodynamic, and cellular activation parameters to test, for example, whether loss of competitive equilibrium between hemispheres (Kinsbourne, 1977), altered gradients of information flow in local circuits (Nakayama, Yamagata, & Hoshi, 2016), changes in preferred directional tuning at the level of individual neurons, or the disruption of cholinergic, adrenergic and dopaminergic neurotransmission (Luvizutto et al., 2015) may be more likely to account for symptoms. These method may be fruitful to co-integrate with lesion-symptom studies (Karnath, Sperber, & Rorden, 2018), once large groups of well-characterized spatial neglect patients with Aiming spatial neglect are available. Thus, researchers studying spatial neglect rehabilitation should consider 1) dividing patients by presence or absence of frontal/striatal injury, or frontal brain disconnection 2) examining pre/post neurophysiologic parameters such as motor or sensory evoked potentials, brain activation, and brain vascular dynamics (Boukrina et al., 2019).

4.d.3. Prism adaptation treatment responders have frontal injury.

A prior study suggested that patients with Aiming spatial neglect at baseline experienced the greatest benefit from prism adaptation treatment, as compared with patients who lacked Aiming bias at baseline—the latter group did not improve after prism adaptation therapy (Goedert et al., 2014). After 10 sessions of prism adaptation therapy over two weeks, we assessed spatial neglect before treatment, and weekly thereafter over 6 follow up assessments, using two measures. One measure was the Behavioral Inattention Test-conventional (BIT-c; Wilson et al., 1987), which we demonstrated is correlated with spatial Where, perceptual-attentional bias (Goedert et al., 2012). The other measure was the Catherine Bergego Scale, which we previously reported accounts for variance in functional performance independent of the BIT-c. The CBS, in a principal components analysis, also contributed to variance in test performance potentially related to motor-exploratory, rather than perceptual-attentional spatial bias (Goedert et al., 2012). Separating patients based on the presence of Aiming spatial neglect did, indeed, predict recovery after prism adaptation therapy (Goedert et al., 2014).

Our discovery that Aiming spatial neglect predicted functional recovery after receiving prism adaptation treatment led us to pursue a neuroanatomic predictor of prism adaptation response. We reasoned that neuroanatomic biomarkers may be a more feasible basis for spatial neglect treatment assignment than behavioral assessment, if the current clinical process is insensitive to spatial neglect features that predict response to prism adaptation treatment, since most patients receive brain imaging during clinical care.. In the previous sections, we discussed the evidence that either frontal cortical injury, or disruption and disconnection of frontal-subcortical networks, may cause Aiming spatial neglect. Thus, we performed two studies of prism adaptation treatment for spatial neglect in which right brain stroke patients were screened for simple presence or absence of lesions anywhere in the frontal cortex (Chen et al., 2014; Goedert et al., 2020). In both of these studies, patients with frontal lesions responded better to prism adaptation treatment based on functional performance assessment with the CBS. The second study (Goedert et al., 2020) included control patients who did not receive prism adaptation therapy, to account for possible better recovery in frontal lobe-lesioned patients. Although both of these studies were small, and their results need to be confirmed before they can be broadly generalized, taken together they strongly suggest there is better response to prism adaptation therapy after frontal lobe injury.

Conclusions.

Patient lesion location can clarify who responds best to prism adaptation treatment for spatial neglect.

In this paper, we review information on Aiming spatial neglect, and present preliminary data supporting assessment of Aiming spatial neglect as a marker for spatial neglect treatment response. We described Aiming spatial neglect, assessed with a computerized line bisection task measuring directional hypokinesia, which has proven value to predict which patients with spatial neglect will improve in their daily life function after receiving early spatial neglect rehabilitation with prism adaptation treatment, and presented preliminary evidence that bromocriptine medication may also specifically alter Aiming spatial neglect during functional task performance. The presence or absence of Aiming spatial neglect on this task, and the presence or absence of brain lesions associated with impaired spatial Aiming networks, both predict improvements in daily life function after prism adaptation treatment. Although this may seem surprising initially, we urge the reader to consider literature on dopaminergic depletion and spatial Aiming in animals, and new data demonstrating that patients with isolated spatial motor Aiming bias may be the group (stratum) responsive to dopaminergic medication intervention.

In future studies of prism adaptation treatment, and other spatial neglect treatments, the very extensive literature on the neuroanatomy and neurophysiology of spatial Aiming bias leads us to examine frontal- and subcortical-connected systems. Reporting whether frontal-subcortical connections have been disrupted in post-stroke neglect, is critical in future prism adaptation studies. If negative studies of prism adaptation treatment for spatial neglect continue to be published, without including any information on patient lesion location, we cannot leverage information from the literature to understand what neuroanatomic-behavioral associations mark the best potential for functional recovery after receiving prism adaptation treatment. In addition, we need to use validated outcome measures in studies of Aiming spatial neglect and spatial neglect treatment. The computerized line bisection task has proven value to predict which patients with spatial neglect will improve in their daily life function after receiving early spatial neglect rehabilitation with prism adaptation treatment, and eventually other tasks, both impairment tasks and functional activities tasks, should be used to assess directional hypokinesia, hemispatial hypokinesia and limb hypokinesia. However, the relationship of the new tasks to the currently-available and validated computerized task assessing Aiming spatial neglect, as well as their relationship to functional outcomes after treatment, will need to be developed.

As described above, we urge future studies examining dopaminergic agents for spatial neglect to consider preferentially enrolling patients with Aiming spatial neglect using the validated computerized task, or to enroll enough participants to perform a three-factor assessment (response, Aiming spatial bias, and brain lesion location). Responders to dopaminergic therapy in previous studies might have been predicted with this procedure (Gorgoraptis et al., 2012).

In spatial neglect treatment studies with prism adaptation treatment, outcomes need to include functional performance to detect improved Aiming spatial neglect.

A second area of major concern is outcomes for spatial neglect rehabilitation studies. Whether outcome measures are biological, cognitive, or motor performance-based, it is important to use disability-relevant functional performance assessment in patients with spatial neglect. This means that measures should be validated to detect functional disability, and not stroke or brain injury. In both of our studies showing stratification of spatial Aiming bias and frontal cortical lesions, the CBS was sensitive to detect differences in patient response, but the BIT-c could not dissociate responsive from unresponsive patients (Goedert et al., 2014; Goedert et al., 2020). Although it may seem obvious, researchers should use outcome measures intended to detect the mechanistic changes their treatments are known to induce. The CBS, as a functional motor performance variable, is likely to be the most appropriate functional disability measure for spatial neglect rehabilitation studies, because of its ability to capture an additional, independent 9% of variability in functional disability over paper-and-pencil tests (Goedert et al., 2012). Also, studies of prism adaptation treatment, medications that stimulate dopaminergic and noradrenergic systems (Luaute et al., 2018), and other spatial Aiming-active interventions, should assess the specific effect of the intervention on laboratory measures on the computerized line bisection assessment of Aiming bias (Garza et al., 2008). As we noted above, this task has been validated by co-occurrence of motor-intentional deficits, and its predictive value in identifying people with spatial neglect whose functional performance improves with early rehabilitation. However, as new tasks are developed assessing other components of Aiming spatial neglect (hemispatial hypokinesia, limb hypokinesia), whether these are laboratory-based or functional performance tasks, it will be important to assess their convergent validity with relation to the computerized line bisection task, and their ability to predict response to rehabilitation. This will not only help identify new ways of predicting and measuring real-life improvement; it will help to clarify how closely related the different manifestations of Aiming spatial neglect may be to each other.

Future studies.

More information needs to be collected on spatial-motor Aiming in 1) animal models 2) human observational studies of spontaneous recovery 3) interventions with different proposed degrees of spatial-motor Aiming influence, potentially as this interacts with different baseline degrees of Aiming spatial neglect, and even combinations of interventions, and continuing efforts to provide overall summary of these efforts would be desirable.

Could the relationship between spatial Aiming and functional performance be mutual: could spatial retraining actually improve adaptive movement and hemiparesis? Spatial retraining is not currently included in routine stroke rehabilitation; its potential value to activate spatial Aiming and motor brain systems, and stimulate motor learning, has not been fully explored. Recently, we (Barrett and Muzaffar, 2014) proposed spatial cognitive treatments might promote better three-dimensional perceptual-motor integration, even when stroke patients have no symptoms of spatial neglect. We advocated the potential for spatial retraining to promote motor recovery: it could improve dexterity and movement coordination, increase strength, and help develop adaptive body movements during ambulation, transfers, self-care, and other functional activities.

Since the neurobiology of brain plasticity is already understood to be a core issue in motor and language recovery after stroke (Small, Buccino and Solokin, 2013), we must also consider how new ways of measuring brain repair can contribute to studying spatial Aiming, including genetic, endocrine, and immune factors. As the paradigm for spatial neglect rehabilitation research broadens to include motor-related brain systems, we look forward to future research that will also help us understand the relevance of biological patient characteristics such as Aiming spatial neglect. We also look forward to studies that evaluate changes in Aiming spatial neglect in parallel with appropriate, disability-relevant outcomes, to assess the therapeutic changes resulting from spatial neglect rehabilitation.