Introduction

Idiopathic blepharospasm (BSP) is an adult-onset focal dystonia that manifests more frequently in women and has a peak age at onset in the fifth to sixth decade [1]. BSP is typically characterized by orbicularis oculi muscle spasms that are usually bilateral, synchronous, and symmetric [2, 3]. Dystonic spasms may be phenomenologically heterogeneous, with either brief or prolonged spasms and narrowing or closure of the eyelids [4]. In addition to spasms, BSP patients may manifest a spectrum of additional signs/symptoms, including sensory symptoms in the eyes [5], increased spontaneous blink rate [6], sensory tricks (stretching, massaging, or touching the eyebrow, the eyelid, or the forehead) transiently improving eyelid spasm [7], apraxia of eyelid opening [8], and dystonia and tremor in other body parts [9]. BSP patients may also have psychiatric disturbances, most frequently depression and obsessive compulsive disorder, and mild cognitive disturbances affecting executive function, with impairment in working memory, processing speed, visual motor ability, and short-term memory [10, 11].

Conventional imaging studies and autopsy findings failed to identify structural brain lesions in idiopathic BSP. Nevertheless, a pattern of functional abnormalities has been established by neurophysiological investigations, including loss of inhibition at different levels of the central nervous system, maladaptive plasticity, and abnormal sensory processing. These abnormalities were traditionally interpreted as the consequence of a primary basal ganglia disturbance. Recent evidence, however, points to anatomical and functional involvement of several brain regions besides the basal ganglia. A summary of individual case reports addressing secondary cranial dystonia and new neuroimaging tools allowing measurement and examination of functional interactions and connectivity among brain regions provided support to the emerging hypothesis of BSP as a network disorder.

The aim of this paper was to review available evidence supporting BSP as a network disorder and to understand whether the proposed network model fits the well-known neurophysiological abnormalities described in BSP.

The network model

Evidence from acquired BSP

Acquired BSP has been associated with structural lesions in several brain regions including not only the basal ganglia but also the thalamus, the brainstem, the cerebellum, and the cortex [12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52] (Table 1). Brain lesions can be focal or more widespread and include ischemia or stroke, tumors, demyelinating lesions, and other pathologies. The onset of BSP can be temporally related to the occurrence of the presumed causative lesion or be delayed by months or even years [29,30,31]. Symptoms can be persistent needing chronic treatment with botulinum toxin or remit spontaneously over a variable period of time or after the treatment of the underlying disease process [15, 16, 25, 27]. Acquired BSP may present as isolated focal dystonia, be part of a segmental or multifocal dystonia, or be associated with other neurological symptoms [12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52].

Table 1 Reports on acquired blepharospasm
Full size table

Evidence from non-conventional magnetic resonance imaging studies

Non-conventional imaging investigations performed in BSP patients include voxel-based morphometry (VBM) studies, diffusion tensor imaging (DTI) studies, and functional magnetic resonance imaging (fMRI) studies.

VBM studies performed in patients with idiopathic BSP showed gray matter changes in the caudate/putamen, thalamus, cerebellum, and cortical/subcortical regions highly relevant to sensory processing and cognitive modulation of motor behavior like the primary sensorimotor cortex and the cingulate gyrus [53,54,55,56,57,58]. Although not tested in all studies, the abnormal changes found in some areas did not correlate with clinical variables such as disease duration or severity (Table 2).

Table 2 Morphometric studies on blepharospasm
Full size table

DTI is a technique that assesses microstructural changes of the nervous tissue through the analysis of water diffusion. Among DTI indices, fractional anisotropy (FA) quantifies directionality of water diffusivity, thus providing a measure of axonal integrity, while mean diffusivity is a measure of cellular and membrane density [59, 60]. No DTI change was observed in three studies comparing 5 to 16 patients with healthy controls [58, 61, 62]. However, a more recent study on 31 BSP patients showed decrease of FA in the white matter of the left anterior lobe of the cerebellum and in the right precuneus of the parietal lobe, and increased diffusivity in the right lentiform nucleus, thalamus, and insula. FA abnormalities significantly correlated with BSP severity and duration. By contrast, increased diffusivity changes lacked any correlation with severity/duration of BSP [63].

Functional magnetic resonance imaging (fMRI) is a tool that studies brain function by exploiting the blood oxygen level-dependent (BOLD) signal changes, i.e., the signal changes due to the increased levels of deoxy-hemoglobin following the activation of brain areas. Signal can be measured while performing specific tasks (task-dependent fMRI) or in the absence of a task or stimulus (resting state-fMRI (rs-fMRI)) [64, 65].

In BSP, task-dependent fMRI studies investigated patients during spontaneous or voluntary blinking [66] and spontaneous spasm [67] with evidence of abnormal activation in subcortical regions and in various cortical areas, including visual and motor cortex. When a task not related to dystonia was applied, abnormalities were found in the basal ganglia and thalamus [68] (Table 3).

Table 3 Task-dependent fMRI studies on blepharospasm
Full size table

In rs-fMRI, different analytic methodologies can be applied, including the amplitude of low frequency fluctuations (ALFF), the regional homogeneity analysis (ReHo), model-free methods like the seed-based analysis and the independent component analysis (ICA), the graph theoretical analysis, and the voxel-mirrored homotopic connectivity [65, 69, 70] (Table 4).

Table 4 Resting-state fMRI studies on blepharospasm
Full size table

ALFF and ReHo methodologies explore the function of and connectivity within specific pre-established brain regions. In BSP, abnormal values of ALFF and ReHo were found in various cortical and subcortical regions consistent with dysfunctions in multiple neural networks [71, 72, 76, 77]. In the studies by Yang [72] and Ni [76], some of these changes correlated with BSP severity, a finding not confirmed in the later investigation of Jiang [77].

Seed-based analysis and ICA assess functional connectivity (FC) of spatially distinct brain areas through the identification of their synchronous BOLD fluctuations at rest. The analysis of FC identified different resting-state networks in the normal brain, including the default mode network (posterior cingulate cortex, medial prefrontal cortex, and lateral parietal cortex), the salience network (dorsal anterior cingulate cortex, bilateral insula, and pre-supplementary motor area), the basal ganglia network, sensorimotor network (within and between left and right sensory and motor cortices), the dorsal attentional network, and the frontoparietal network (lateral prefrontal cortex and inferior parietal lobule) [65, 69]. In BSP patients, seed-based analysis and ICA showed dysfunctions within the sensory motor network, the salience network, the default mode network, and the right frontoparietal network [71, 73,74,75].

The graph theoretical analysis is an analytic methodology that measures brain functional organization on a large-scale level and allows examination of the whole brain connectivity patterns. [65]. By this approach, abnormal network architecture at large-scale level was found in BSP patients along with the detection of sensorimotor and frontoparietal networks’ FC alteration [73].

Finally, the voxel-mirrored homotopic connectivity, a method of rs-fMRI designed to compare the interhemispheric rs-fMRI FC revealed enhanced homotopic coordination in the brain regions associated with sensory integration networks and default-mode network [70].

Overview of neuroimaging studies

Reports on acquired BSP have limitations. First, the structural abnormalities associated with BSP lack specificity because it may also be found in non-dystonic people; second, we cannot exclude a physiopathological role of microstructural defects or functional disturbances arising from apparently normal brain regions; finally, a temporal relationship between lesion development and emergence of symptoms is not always evident. Therefore, it is sometimes difficult to rule out the possibility of an idiopathic BSP with coincidental lesions [52].

Nevertheless, it is of interest that the microstructural alterations found in idiopathic BSP on DTI and VBM investigations largely affect brain regions associated with acquired BSP. The possibility that a structural lesion within a neural network induces functional changes leading to the appearance of dystonic symptoms still remains a plausible physiopathological mechanism. The consistency of the findings between idiopathic and acquired BSP strongly supports the hypothesis that many brain regions are involved in the physiopathology of BSP. Supporting this view, information from functional imaging studies indicated that BSP patients may have both abnormal intraregional brain activities and interregional dysfunctional connectivities and supported derangements in communication among frontal and parietal cortical areas, the basal ganglia, the thalamus, and the cerebellum.

Another issue to be considered is the lack of correlation between severity/duration of BSP and the abnormalities found in some regions of the brain. Although the correlation between clinical variables and imaging findings was not tested systematically, this observation would suggest that the abnormalities found in multiple brain regions not limited to the basal ganglia may be trait alterations in BSP patients.

In conclusion, information provided by imaging studies supports the hypothesis that BSP may arise from a disordered brain network [78].

Neurophysiological abnormalities

Several neurophysiological abnormalities affecting sensorimotor pathways at different levels of the central nervous system have been described in BSP (Table 5).

Table 5 Neurophysiological studies on blepharospasm
Full size table

The earliest detected abnormality was the decreased inhibition of the R2 response of the blink reflex by paired electrical stimulation of the supraorbital nerve [79,80,81, 84, 85] that relies on brainstem oligosynaptic circuits [87]. As botulinum toxin treatment leaves the blink reflex recovery cycle unchanged but significantly improves OO muscle spasms, [82, 83] the blink reflex abnormalities in BSP are likely to reflect a pathophysiological mechanism rather than dystonic activity in the OO muscle. The enhanced excitability of interneurons in the brainstem extends outside the blink reflex to include an enhanced recovery curve of the silent period 2 of the masseter inhibitory reflex [85, 92], and the trigemino-facial circuits to include an abnormal auditory startle reaction [88], an abnormal trigemino-sternocleidomastoid reflex, [84] and an abnormal somatosensory pre-pulse modulation of the blink reflex by peripheral stimuli [81, 86]. Brainstem interneuron hyperexcitability secondary to a suprasegmental dysfunction may explain the enhancement of the recovery curves of the R2 response of the blink reflex as well as the abnormal changes observed in other brainstem reflex responses [87, 88, 92]. The basal ganglia would influence the blink reflex circuit through two main routes. First, basal ganglia projections via thalamus to cortex might change the activity in descending cortico-brainstem pathways controlling blink reflex excitability. Alternatively, there could be a route via projections to either pedunculopontine nucleus (PPN), or most likely the superior colliculus (SC); in turn, outputs from the SC could project to the raphe magnus and hence to interneurons in the trigeminal sensory nucleus [93,94,95].

Transcranial magnetic stimulation (TMS) showed abnormal excitability of the primary motor cortex (M1) in BSP, as demonstrated by reduced short-interval intracortical inhibition (SICI) in the hand muscles and by reduced duration of the cortical silent period in the cranial muscles of such patients [89, 90]. Using paired associative stimulation, a technique that investigates cortical plasticity, Quartarone and colleagues [91] observed that the plasticity of cortical motor areas is increased in the hand muscles of patients with BSP. Although the aforementioned TMS were traditionally attributed to a basal ganglia dysfunction, deep brain stimulation (DBS) studies performed in patients with a variety of movement disorders have shown that stimulation of the subthalamic nucleus (STN) restores SICI in patients with Parkinson’s disease (PD) [96]; stimulation of the ventralis intermedius nucleus of thalamus (VIM) may enhance M1 excitability to TMS in patients with essential tremor [97]; stimulation of the STN may improve silent period changes in patients with PD [98]; and GPI stimulation may modulate M1 excitability in patients with dystonia [99]. Finally, electrical or magnetic stimulation of the cerebellum can modulate excitatory and inhibitory M1 circuits [100,101,102,103].

Earlier studies tested the sensory system in dystonic patients by assessing somatosensory-evoked potentials (SEPs) that is electrical potentials generated in sensory pathways at peripheral, spinal, subcortical, and cortical levels of the nervous system [104]. SEPs may be used to assess sensory integration in the time domain by applying the paired-pulse paradigm; in normal subjects, a preceding (conditioning) stimulus induces suppression of SEP amplitude evoked by the following (test) stimulus [105]. In patients with dystonia, this inhibitory effect is impaired [106,107,108]. Supporting this view, several studies on patients with various forms of focal dystonia including BSP have reported increased somatosensory temporal discrimination thresholds (STDT), the shortest time interval needed to discriminate two consecutively applied tactile stimuli [109,110,111,112]. Increased STDT values are present in body parts that may be affected or unaffected by dystonia and do not improve after botulinum toxin [112, 113]. Because altered STDT has been observed in BSP and in other focal dystonias, does not correlate with clinical severity, and has been also observed in unaffected first-degree relatives of patients, the abnormal temporal processing of sensory information is likely to enhance the permissive environment that predisposes patients to the development of dystonia [106, 114,115,116,117,118]. Supporting this view, STDT abnormalities have been reported in patients with increased blinking as prodromal phase of BSP [119, 120].GABA-mediated mechanisms of inhibition in the somatosensory cortex S1 as well as a dopamine-mediated subcortical network, involving the basal ganglia, thalamus, and superior colliculus, probably contribute to the mechanisms underlying STDT [111, 121, 122].

Overview of neurophysiological studies

Neurophysiological investigations in idiopathic BSP found abnormal functional changes that were consistent with loss of inhibition at different levels of the nervous system. Some of these studies may have been limited by the small size of the study sample and the lack of attention to the possible correlations between neurophysiological changes and clinical variables. Furthermore, several neurophysiological abnormalities could also be found in movement disorders other than dystonia [123,124,125,126]. Nevertheless, loss of inhibition is a well-defined functional trait of dystonia [127] that could also be detected in unaffected body regions of dystonic patients or in non-manifesting DYT1 mutation carriers, thus raising the possibility that these changes are an endophenotype of dystonia [128,129,130].

Discussion

Although BSP is traditionally considered a basal ganglia disorder, accumulating evidence from structural and functional investigations points to the anatomical and functional involvement of several brain regions. This has led to the hypothesis that BSP may arise from a disordered brain network [78]. Considering the remarkable clinical heterogeneity of motor and non-motor manifestations characterizing BSP, a related heterogeneity of the underlying anatomical substrates would not be surprising.

Available information makes it highly likely that the primary defect in BSP lies somewhere in a network connecting the basal ganglia, thalamus, frontal and parietal cortices, cerebellum, and brainstem [78]. Likewise, similar findings from clinical reports and non-conventional imaging studies also support a derangement of brain networks in other focal dystonias like cervical dystonia and upper limb dystonia [78, 131]. If the network model can also explain the well-established neurophysiological abnormalities that are present in BSP and were previously attributed to a basal ganglia disturbance, then this would enhance the likelihood of the network model itself.

In patients with BSP, neurophysiological studies highlighted loss of inhibition at cortical and sublevels (cortical and subcortical) of the central nervous system. Theoretically, changes of excitability in restricted brain areas like cortical M1, brainstem, and cortical S1 may depend on intrinsic disorders in these areas and/or may reflect abnormal influences from distant brain structures. The latter hypothesis is supported by several evidences: DBS studies have shown that stimulation of the STN, GPi, or VIM modulates M1 excitability [96,97,98,99]; excitatory and inhibitory M1 circuits are also affected by electrical or magnetic stimulation of the cerebellum [100,101,102,103]; and GPi stimulation increases R2 inhibition of the blink reflex recovery curve in dystonic patients [132]. Therefore, changes in cortical/brainstem excitability may reflect dysfunctions possibly involving the basal ganglia-thalamo-cortical projections and the cerebello-thalamo-cortical projections. Likewise, cortical S1 abnormalities underlying STDT disturbances may develop secondarily to abnormal connectivity from subcortical projections from the basal ganglia, thalamus, and superior colliculus rather than occurring independently.

If impaired inhibition at cortical/brainstem levels and tactile information processing abnormalities can be interpreted as dysfunctional connectivity between subcortical networks and M1/S1 cortices, then neurophysiological abnormalities carried by BSP patients would be coherent with the pathophysiological model positing that BSP arise from a disordered network connecting the basal ganglia, thalamus, frontal and parietal cortical areas, cerebellum, and brainstem. In this network organization, the thalamus would be a central dysfunctional hub because it not only gates bottom-up and top-down streams of sensory information directed to and from the cortex but also integrates sensory inputs of different modalities with output from the basal ganglia and cerebellum [133]. The cerebellum might be considered an additional dysfunctional hub in the network model of dystonia, as suggested for cervical dystonia and focal hand dystonia by dysfunction of cerebello-thalamo-cortical circuits [134,135,136] and impaired cerebellum-dependent associative learning explored by the classical eyeblink conditioning paradigm (EBCC), a Pavlovian learning protocol integrated at the level of Purkinje cells and deep cerebellar nuclei [137,138,139]. However, there is no neurophysiological evidence as yet of cerebellar involvement in idiopathic BSP. Future studies designed to assess the EBCC and the functional interactions between cerebellum and M1 in BSP patients may clarify whether different forms of focal dystonia share the same dysfunctional circuits.

In conclusion, evidence from structural and functional investigations supports the hypothesis that BSP could arise from a disordered brain network. This model seems also to fit the neurophysiological abnormalities so far reported in idiopathic BSP. However, several questions still remain open. Further studies are needed to better understand which is the core of the dysfunctional network and which are sites that can be permissive; which abnormalities play a causative role and which are compensatory phenomena; which factors may cause BSP to fully clinically express; how structural and functional abnormalities correlate with clinical symptoms and disease evolution; and whether the spectrum of non-motor symptoms characterizing BSP is in some way related to the abnormal network of brain structures thought to contribute to the condition.