|Home | About | Journals | Submit | Contact Us | Français|
The insula, a ‘cortical hub’ buried within the lateral sulcus, is involved in a number of processes including goal-directed cognition, conscious awareness, autonomic regulation, interoception and somatosensation. While some of these processes are well known in the clinical presentation of migraine (i.e., autonomic and somatosensory alterations), other more complex behaviors in migraine, such as conscious awareness and error detection, are less well described. Since the insula processes and relays afferent inputs from brain areas involved in these functions to areas involved in higher cortical function such as frontal, temporal and parietal regions, it may be implicated as a brain region that translates the signals of altered internal milieu in migraine, along with other chronic pain conditions, through the insula into complex behaviors. Here we review how the insula function and structure is altered in migraine. As a brain region of a number of brain functions, it may serve as a model to study new potential clinical perspectives for migraine treatment.
The focus of brain systems biology has trended towards integrated networks, which comprise a jigsaw of interacting brain regions. One region, the insula, located bilaterally in the lateral sulcus, plays a complex role in emotion, homeostasis (including error detection), autonomic function, sensation, salience, and awareness (Nieuwenhuys, 2012). It is also known to be involved in specific behaviors related to disease conditions such as migraine. Migraine attacks involve a wide range of sensory, emotional, and cognitive symptomology. It is thus conceivable that the insula may serve as a cortical hub, processing many of the complex sensory and emotional aspects known to be present in the migraine condition.
In this review, we explore known functional components of the insula and their potential role in migraine. Specifically, we first summarize Anatomical Aspects of the Insula in the first section and then briefly outline Functional Divisions of the Insula. Secondly, we discuss The Functional Insula in Migraine, on how the insula that may contribute to some of the observed sensory, physiological, psychological, and cognitive changes in migraine. In the final section, Insula and Migraine Therapy, we review data that support modulation of insula function in therapies used in migraine treatment. The review attempts to provide an overview of insular function in regards to migraine. Clearly, migraine is not “an insula disease”. However, since the insula is involved in complex behaviors associated with the migraine state, both ictally (the period of the migraine attack) and interictally (the period between migraine attacks), the insula could paly an important role in integrating many of the dynamic processes know to be involved in migraine (e.g., sensory, autonomic, cognitive). In addition, while attempting to provide a review of current information regarding this region’s involvement in migraine, it is a brain region involved in other conditions (i.e. fibromyalgia, chronic lower back pain, etc.) and needs to be considered in this context.
The insula is a brain region located within the Sylvian fissure in the fronto-parietal and temporal opercula (Guenot et al., 2004). Figure 1A shows the gross anatomy of the insula. The insula consists of 5 – 7 lobes in humans, which are divided into anterior and posterior portions by the central insular sulcus. Further anatomical divisions have been described elsewhere (Ture et al., 1999). All insular gyri are interconnected, except the anterior and posterior short gyrus (Almashaikhi et al., 2014). Recent studies in humans using tractography (the process of tracing white matter tracts using MRI) reveals that the anterior insula is predominantly connected to limbic and paralimbic brain regions as well as the anterior parts of the inferior frontal gyrus, while the posterior insula is connected with parietal and posterior temporal cortices (Cerliani et al., 2012). These findings correlated with connections have also been reported in anatomical tract tracing studies in primates (reviewed by (Augustine, 1996)). The insula also has various connections with other cortical areas that include the visceral sensory area, visceral motor area, motor association area, vestibular area, as well as subcortical connections to the amygdala (Sah et al., 2003) and thalamus (Mufson and Mesulam, 1984). Accordingly, the insula is highly interconnected throughout the brain, thus providing the underpinnings of its involvement in a wide range of functions. Figure 1B shows horizontal and coronal sections through the insula providing a guide for the relative locations of the activations or gray matter changes shown in Figures 3 and and44.
In addition to the complex connectivity, the neuronal population of the insula varies across the different subregions of the brain structure: from granular neocortex in the posterior-dorsal insula to agranular neocortex in the anterior-ventral insula (Bauernfeind et al., 2013). Specialized neurons, called von Economo Neurons (VENs), have been found in the insula (Allman et al., 2011; Evrard et al., 2012), specifically in the agranular frontoinsular (rostro-ventral) part (Bauernfeind et al., 2013). In humans, VENs are observed in the 36th week of gestation, are rare at birth, and increase in number during the first 8 months after birth (Allman et al., 2011). As they develop, there is a laterality with more found in the right insula (Allman et al., 2010), which shows dominance for a number of functions (see below). It has been therefore suggested that they play a critical role in complex characteristics like self-awareness and social cognition (Critchley and Seth, 2012). Differences in biochemical specificity that have been reported in VENs in humans vs. other hominids may relate to evolutionary development of specific processing such as interoception (Stimpson et al., 2011).
The insula can also be functionally divided into anterior and posterior parts and exhibit functional divisions involved in somatosensory, autonomic, interoceptive, salience, and cognitive processing (see Figure 1C). The variation in function of the human insula (Cloutman et al., 2012) seems to correlate with structural connectivity noted above (Cerliani et al., 2012). Functionally, the anterior insula shows connections within the middle and inferior temporal cortex and anterior cingulate cortex, and is primarily related to limbic regions, which play a role in emotional modulations (Cauda et al., 2011). Furthermore, the anterior insula may also serve as a hub that integrates interactions with large-scale brain networks (Menon and Uddin, 2010) as it forms an integral part of the salience network (Damoiseaux et al., 2006), thus potentially incorporating behavioral responses with internal or external salient stimuli (Borsook et al., 2013b). The posterior region is more closely connected to the premotor, sensorimotor, supplementary motor and middle-posterior cingulate cortices, indicating a role for the posterior insula in sensorimotor integration (Cauda et al., 2011). Finally, anterior and posterior divisions of the insula are also functionally connected (Cauda et al., 2011) (see Figure 1D). Further details of the functional role of the insula are reviewed elsewhere (Afif and Mertens, 2010; Bauernfeind et al., 2013). It should also be noted that other authors have offered a slightly different version of sub-regional functions of the insula (Cauda et al., 2011; Deen et al., 2011; Kurth et al., 2010).
Although multiple functional features of the insula have been described, an understanding of the details of the functional anatomy of the insula is still elusive. There are some overarching features such as converting physiological states into emotions related to feeling and a sense of being. However, based on surgical, electrophysiological and imaging studies, different anatomical regions may have specific functional domains (Cauda et al., 2014; Dupont et al., 2003). Direct stimulation of the anterior insula in humans elicits very few reportable responses, whereas stimulation of the mid and posterior parts of the insula results in gustatory and somatosensory symptoms (Stephani et al., 2011). Painful sensations have also been reported following stimulation of the posterior insula (Mazzola et al., 2009), with more specificity noted in the upper posterior part of the human insular cortex and right-sided lateralization (Ostrowsky et al., 2002). This latter finding has been confirmed by functional magnetic resonance imaging (fMRI) studies showing that the insula is somatotopically organized (Brooks et al., 2005). Interestingly, the integration of visceral and somatic inputs into more complex outputs from the insula may relate to intra-insular functional connectivity, as significant connectivity between anterior and posterior sub-regions has been observed during neurosurgical evaluation in patients with refractory epilepsy (Almashaikhi et al., 2014).
Figure 2 summarizes afferent and efferent pathways involved in basic functions of the insula (viz., somatosensory, visceroception, cognition, etc.). These connections form the basis for many of the altered functions observed in migraine. Inputs include pain through trigemino-vascular afferents, dizziness through vestibular afferents, and visceral connections that may provide a basis for abdominal symptoms and autonomic changes. Insular outputs to subcortical and brainstem regions that contribute to cortical control of autonomic function, the cingulate cortex contributing to interoceptive responses, and to the frontal lobe that is related to cognitive processing.
Based on its functional role and widespread connectivity with cortical and subcortical brain regions the insula has been proposed to play a significant role in neuropsychiatric disorders (Nagai et al., 2007) including mood and anxiety disorders (Buse et al., 2013; Sheftell and Atlas, 2002), temporal lobe epilepsy (Ostrowsky et al., 2000), craving and drug seeking in addiction (Naqvi et al., 2014), pain (Mazzola et al., 2009; Ostrowsky et al., 2002), and non-motor symptoms of Parkinson’s disease (Christopher et al., 2014). In fact, damage of the insula affects a number of important behaviors and perceptions including individuals’ sense of being, apathy or tiredness (Manes et al., 1999), neglect (Manes et al., 1999), temperature perception (Baier et al., 2014), dysarthria (Baier et al., 2011), auditory agnosia (Bamiou et al., 2003), processing of empathy (Gu et al., 2012), and drug-craving (Garavan, 2010).
As noted above, many of the putative functions ascribed to the insula also appear as symptoms in migraine so studying insular function and dysfunction may lead to further understanding and treatment. Table 1 lists imaging studies on migraine reporting insular changes (see criteria for PubMed search in table key). Migraine is more than a headache. It involves altered sensory, emotional, and cognitive processes (Charles, 2013). In the following sections, we will explore 4 principle domains of altered function (sensory, physiological, psychological and cognitive) in migraine and the putative involvement of the insula in these functions. Examples of insula activation across processes that can induce changes specific to migraine symptoms are those defined for underling functions in other studies including pain (Henderson et al., 2008), autonomic function (Henderson et al., 2012; James et al., 2013), cognition (Wiech et al., 2010), prediction error (Bossaerts, 2010; Preuschoff et al., 2008), and aversive processes such as disgust (Wright et al., 2004), vestibular function (Fasold et al., 2002) and nausea (Napadow et al., 2013).
In Figure 3A demonstrates the functional changes to the insula observed in migraine subjects compared to healthy controls. There is an increase in insular activity that corresponds to the default mode network (DMN) (Xue et al 2012), a network associated with normal brain function while at rest. Figure 3B displays the correlation between the number of years with migraine and increased functional alterations to the resting state network (RSN) (Schwedt et al 2013), a network synonymous with the DMN. The results of these studies indicate that migraine may not only alter insular activity but over time, these repeated attacks continue to continue to increasingly effect insula function.
The major sensory symptom of migraine is head pain that is severe and debilitating (Silberstein, 2004). The posterior insula has inputs from the thalamus conveying nociceptive signals to the insular cortex (Craig, 2003; Craig et al., 2000). The insula has been called “a multidimensional integration site for pain” (Brooks and Tracey, 2007). Separation of innocuous (e.g., touch, proprioception, innocuous cold or warmth) and noxious stimuli (heat) has been reported in the human insula, as proprioception activates the contralateral mid-insula, innocuous cooling activates mid- and dorsal posterior insular parts, and pain activates the contralateral posterior insula, indicating that the insula may contribute to sensory-discriminative functions (Mazzola et al., 2012). Most recently, data obtained using intracerebrally recorded nociceptive laser evoked potentials (LEPs) from the full extent of the insula suggests that nociceptive input is first processed in the posterior insula, where it is coded in terms of intensity and anatomical location, after which is conveyed to the anterior insula, where the emotional reaction to pain is processed (Frot et al., 2014). Furthermore, both heat and pinprick stimuli demonstrated contralateral anterior and posterior insula activation (Baumgartner et al., 2010), further supporting the notion that the insula may be involved in sensory integration in pain. In addition, the pain inputs may be further incorporated into complex networks, such as the salience network, through the anterior insula (Wiech et al., 2010). Painful stimulation also leads to an increase in insula to periaqueductal grey functional connectivity (Linnman et al., 2012). Functional connectivity data for anterior and posterior insular regions activated by noxious and non-noxious stimuli have led to further segregation of pain processing, showing that the anterior insula is more strongly functionally connected to affective and cognitive regions, whereas the posterior insula has strong connectivity with sensory-discriminative processing regions (Peltz et al., 2011). The frequent activations of the anterior insula observed in pain fMRI studies may be interpreted as “heightened alertness of either stimulus- or task-driven origin, or both” that integrates internal and external stressors to maintain allostasis (Sterzer and Kleinschmidt, 2010).
Given that the insula appears to be a brain hub or convergence point for afferent inputs (predominantly nociceptive) and emotional processing, one would expect similar activations in pain conditions such as migraine (notwithstanding that other sensory information may flow in parallel). Cortical projections from trigeminovascular neurons in the thalamus have been described to the insula – as well as other cortical areas (Noseda et al., 2011). In patients with insula lesions, acute experimental noxious stimuli produce higher pain intensity ratings and an increased level of responses in the primary somatosensory cortex (S1) compared to healthy controls, suggesting that the insula is involved in “tuning cortical regions to appropriately use previous cognitive information during afferent processing” (Starr et al., 2009). Furthermore, a case of an epileptic focus in the posterior insula has been described, which produced pain (as opposed to inducing post-seizure related headaches) during a seizure and, in addition, stimulation of the same area elicited by pain (Isnard et al., 2011). Mapping of seizure-induced pain in the insula has shown predominantly in the sensory region of the posterior insula (Ostrowsky et al., 2002) supporting the notion that the posterior insula is involved in pain perception. Figure 3C demonstrates how high frequency (HF) migraineurs show decreased activation to painful stimuli compared with low frequency (LF) migraineurs (Maleki et al., 2012a). Additional findings of decreased cortical thickness of the insula in the HF group indicate a dynamic nature of this region with pain. This suggests that insular dysfunction may result form increased migraine frequency (Maleki et al., 2012a). Other insular changes in migraine are illustrated in Figure 3. The dynamic functional and morphological changes in the of the region include integrative process whereby sensory information is transferred to the anterior insula involved in emotional processing of pain (Frot et al., 2014). In this way repetitive migraine attacks may alter emotional processing through regions such as the insula.
Other sensory alterations in migraine include vestibular dysfunction, which in turn contributes to symptoms of nausea and vomiting. Migraine may present a forme fruste ‘vestibular migraine’ condition (Furman et al., 2013); generally, many migraineurs have symptoms that are vestibular in nature. Small lesions of the insula may present with vertigo (Papacostas et al., 2006) and posterior insula strokes may coincide with both pseudothalamic sensory and vestibular-like syndromes (Cereda et al., 2002). Vestibular neurons in the insula have been found in monkeys (Grusser et al., 1990) that were affected by neck and visual responses. One of the reasons patients with migraine typically obtain relief from laying down may therefore relate to limiting neck and visual inputs that may contribute to vestibular symptoms. Vestibular symptoms, while more common in children and adolescents (Weisleder and Fife, 2001), are present in some 30 – 50% of migraine patients (Stolte et al., 2014).
Both smell and taste are altered in migraine, however, the most common symptoms relate to osmophobia (De Carlo et al., 2010; Kelman, 2004). Described as olfactory aura, burning smells, although uncommon, may precede or come on with the headache of some migraine attacks (Coleman et al., 2011). Osmophobia is frequently present in both pediatric and adult migraineurs (De Carlo et al., 2012; Wang et al., 2012; Zanchin et al., 2005). While around 20% of migraineurs have been reported to be anosmic or hyposmic (Hirsch, 1992), positron emission tomography (PET) studies show hypersensitivity to olfactory stimuli (Demarquay et al., 2008) with activations in a number of brain regions such as the temporal lobe, anterior cingulate, and locus coeruleus, but not the insula. Interictal olfactory hypersensitivity is reported (Marmura et al., 2014; Stankewitz and May, 2011) to be present in about 35% of migraineurs (Demarquay et al., 2006) and seems to be a predictor of olfactory triggers for migraine attacks. In addition, during a migraine attack some migraineurs have olfactory symptoms of microsmia or hyposmia (Marmura et al., 2014). During spontaneous migraine attacks, there is increased activation induced by olfactory stimuli (rose odor) in the amygdala and insula (Stankewitz and May, 2011). The perception of taste is integrated in the anterior insula and in patients with migraine may be involved in the aversive response or disgust to food as a potentially protective mechanism concomitant with symptoms of nausea and vomiting. Afferents from viscera terminate in the granular and dysgranular parts of the insula (Stephani et al., 2011). Taste is represented in the insula in both humans (Small, 2010) and rodents (Sewards and Sewards, 2001). Indeed, the insular cortex is the primary cortical region involved in taste and smell (Maffei et al., 2012), which may be part of a visceral response circuit that regulates food cravings and aversions (de Araujo and Simon, 2009). Interestingly, lesions of the insula are known to affect olfaction in humans (Mak et al., 2005; Stevenson et al., 2013), as well as diminished taste recognition and intensity deficits (Pritchard et al., 1999). Seizure-related emesis (ictus emeticus) (Shuper and Goldberg-Stern, 2004) has been reported in patients with lesions in the insula, suggesting a trigger zone for this process (Fiol et al., 1988) involving the spread of abnormal electrical activity through descending insular or limbic circuits (Shuper and Goldberg-Stern, 2004). Persistent nausea is a marker for severe migraine (Lipton et al., 2013). In addition to symptoms of nausea, abdominal pain may be a correlate of migraine (abdominal migraine) in children and adolescents (Carson et al., 2011). Accordingly, these findings confirm the complexity of visceral pathways in migraine, in which the insula may represent a major brain-processing site.
The insula is involved in a wide range of physiological processes that are clinically altered in migraineurs, some of which are discussed below.
In migraineurs, alterations in autonomic function have been well documented. Overall, adult patients with migraine exhibit autonomic nervous system hypofunction, as measured in the interictal period (Shechter et al., 2002). Additionally, we have recently shown alterations in hypothalamic functional connectivity in episodic migraineurs compared with healthy controls (Moulton et al., 2014), thus possibly reflecting abnormal autonomic system. Altered hypothalamic function during migraine (Alstadhaug, 2009) may also suggest that the autonomic nervous system can trigger migraine attacks. The insula is one of many cortical regions involved in autonomic control (Cechetto, 2014; Jones, 2011; Nagai et al., 2010). Stimulation of cardiopulmonary afferents produce activation in the anterior insular cortex through a pathway that originates in the nucleus of the solitary tract and synapses in the parabrachial nucleus and the ventroposterior parvocellular nucleus of the thalamus (Cechetto, 2014) (see Figure 2). Autonomic function in the insula in migraine is still to be evaluated, but based on evidence of pain-related activations of the insula and its role in autonomic regulation (Leone et al., 2006), it is highly likely that altered regulation of autonomic function may also involve changes in insular functions (Geraud and Donnet, 2013). Further evidence has been seen in strokes occurring in the insula that result in increased susceptibility to cardio-autonomic dysfunction (Meyer et al., 2004).
Sleep disruption is reported in migraine patients (Jennum and Jensen, 2002; Sahota and Dexter, 1990). During rapid eye movement (REM) sleep there is a relative activation in a number of brain areas (limbic and paralimbic) including the insula compared with the awake-state as assessed by PET fluoro-deoxy glucose (FDG) imaging, suggesting that involvement of these regions is part of “the integration of neocortical function with basal forebrain-hypothalamic motivational and reward mechanisms” (Nofzinger et al., 1997). In a follow-up study for slow wave sleep (SWS), the same group reported that whole brain metabolism decreased relative to the awake-state, even after controlling for whole brain decreases, but no changes were observed in the insula (Nofzinger et al., 2002). It has however been recently indicated that slow wave function may correlate with the thickness of the insula (Dube 2015 PMID:25995467). Others have also reported the role of the insula in the sleep-wake cycle including REM and non-REM components (Braun et al., 1997). The potential interaction of the insula and migraine may relate to adaptive and restorative changes, since sleep deprivation alters insula function related to decision-making (Venkatraman et al., 2011) and autonomic function (Konishi et al., 2013; Meerlo et al., 2008), including neural circulatory control (Kato et al., 2000). Given the important role of the insula in autonomic function and that alterations in cerebral blood flow may contribute to the headache phase (Asghar et al., 2011), altered sleep may thus diminish resilience in migraineurs. Figure 4B shows another example of the importance and effect of sleep with a decrease in activation in the insula from sleep deprivation.
The role of the insula in interoception was introduced by Craig in the context of pain and sensation relating to mechanisms around ‘how do you feel’ the physiological state of the body (Craig, 2003). While other brain regions may be involved in sense of self (Critchley et al., 2004), the insula plays a key role in connecting these physiological interoceptive processes with feelings (Pollatos et al., 2007). The dorsal posterior insula is involved in interoceptive awareness (Craig, 2003), and more recently, the anterior insula has also been implicated in this process (Zaki et al., 2012). The overall integration of interoception may involve contributions of “discrete modules” (Evrard et al., 2014) and its integration in complex (e.g., salience-related (Borsook et al., 2013b) circuitry and interpretation of sensory inputs to the posterior insula and integration of autonomic control from the anterior insula. Interoception is likely important in migraine since a migraineur’s physiological internal state is different in ictal and interictal states. The physiological state in episodic migraine is manifestly altered in the peri-ictal phase with major sensory and autonomic changes that presumably have effects on interoceptive processing changing ‘feelings, energy, and effort’ and thus the subjective state of being (Craig, 2013).
Compared with males, females have a higher incidence and prevalence of migraine (Silberstein and Lipton, 1993; Stewart et al., 1994b). Sex differences relate to both brain and behavioral changes. Recently, our group reported structural and functional sex differences in the insula in female vs. male migraineurs (Maleki et al., 2012b). Furthermore, another study by our group shows that the insular cortex does not show normal thinning with age (Maleki et al., 2014 In submission). The latter is of interest since in other clinical disorders such as MDD, patients displayed decreased gray matter volumes in the left dorsal anterior insula (Liu et al., 2014). Figure 3D shows sex differences of insula gray matter volume in male and female migraineurs.
The prevalence of migraine decreases significantly after the age of 65 years in both males and females (Victor et al., 2010). Between the ages of 12 – 30 years, the fiber density decreases between the insula and the frontal and parietal cortices but increases between the insula and the temporal cortex (Dennis et al., 2014). Functional implications also relate to insular connectivity in chronic pain states where brain structure and function shift from being adaptive in younger to being maladaptive in older patients (Ceko et al., 2013). The insula changes reported in aging (Foundas et al., 1998) result in functional alterations in pain processing (Tseng et al., 2013).
Migraineurs exhibit significant psychiatric comorbidities (Buse et al., 2013; Sheftell and Atlas, 2002). In addition, migraineurs have altered processing in psychological domains such as mood (Marino et al., 2010), tiredness (Raggi et al., 2012), and disgust and/or unpleasantness of environmental stimuli (Demarquay et al., 2006). The insula has been defined as a ‘limbic integration cortex’ and is putatively involved in emotion. Some of these are discussed in more detail below.
As noted in a recent review, the insula may be involved in anxiety regulation (Paulus and Stein, 2006, 2010). Anxiety-prone subjects show increased activity in the anterior insula (Simmons et al., 2011); additionally, this subgroup showed increased insula activation in response to the anticipation of aversive stimuli (Simmons et al., 2006). In migraineurs, anticipation may contribute to enhancing anxiety-related circuits that include the insula and thus ‘drive’ a more anxious phenotype to expect the onset of the next migraine attack. Interestingly, panic attacks have a high comorbidity with migraine (Smitherman et al., 2013; Stewart et al., 1994a). Physical exercise (e.g., yoga) may modulate the nociceptive/pain afferent input and potentially the emotional reactions, such as anxiety, to the insula resulting in a change in insular brain anatomy and connectivity (Villemure et al., 2013). Related to anxiety is the phenomenon of panic attacks that are comorbid with migraine (Smitherman et al., 2013; Stewart et al., 1994a).
Migraine is a stressor (Borsook et al., 2012; Radat, 2013). Measures of cortisol in migraine patients were found to be higher than control subjects at most times tested (Ziegler et al., 1979). “Safety signals are learned cues that predict stress-free periods whereas behavioral control is the ability to modify a stressor by behavioral actions” and in this way diminish the effects of stressors (Christianson et al., 2008; Christianson et al., 2011). Part of this stress inducing safety response may be through autonomic regulation (Cechetto, 2014). In tracing studies in rats, there is a convergence of autonomic and limbic function that may underlie the interaction of viscera-somatic inputs and behavioral processes (Saper, 1982). Taken together, the insula may be intimately involved in migraine, possibly through (1) integration of safety signals (Christianson et al., 2011), (2) autonomic function (Mosek et al., 1999; Shechter et al., 2002) (see above), and (3) translation of stress (Bossaerts, 2010) into behavioral outcomes (Craig, 2010). As part of the salience network (Borsook et al., 2013b), the region may help determine the resilience to stressors (i.e., repeated migraine attacks).
Neuroimaging studies have shown lateralization of affective pain and other processes (Duerden et al., 2013; Mutschler et al., 2012). Right-sided somatosensory lateralization of insula activation is evident from a number of studies in pain (Brooks et al., 2002; Symonds et al., 2006). In a PET study of spontaneous migraine, right-sided activation in the insula was present (Afridi et al., 2005a) (see Figure 3E). Right-sided lateralization has also been observed in cortical thickness in female migraineurs compared to male migraineurs (Maleki et al., 2014 In submission). In patients with major depressive disorder (MDD), insular hypoactivity is reported during homeostatic shifts, suggesting a deficit in the ability to have a normal response to changes in the environment (Strigo et al., 2010). The undulating onset and offset of migraine attacks may show micro episodes that parallel the MDD group, suggesting that mood swings may manifest in altered right insula processing (Coen et al., 2009).
Fatigue impacts some 70 – 84 % of migraine patients; it may last for days after the attack and is the most frequent psychosocial difficulty in these patients (Raggi et al., 2012). As reports of insular damage show, a putative basis for these changes may result from a functional disconnection with brain structures (e.g., frontal lobe, anterior cingulate cortex) involved with willed motor behavior (Manes et al., 1999).
Disgust is one of basic emotion (Chapman and Anderson, 2012; Toronchuk and Ellis, 2012) that can be elicited by a wide variety of stimuli, both concrete or abstract in nature (e.g., rotten food, immoral persons) (Suzuki, 2010). Imaging studies on disgust have implicated the role of the insula (Phillips et al., 2004; Sprengelmeyer et al., 2011; Suzuki, 2012; Wicker et al., 2003; Wright et al., 2004). Since pain and disgust have been suggested to share some common themes (Kunz et al., 2013), the insula may therefore be responsible for certain processes in the peri-ictal migraine state (e.g., pain).
Neuroaesthetic processing may involve the evaluation and appreciation of valence of sensory processes (Brown et al., 2011). Unpleasant feelings can be induced in migraineurs taking triptans (Kramer et al., 2007). In the latter context, an fMRI study reported that following sumatriptan injection, increased activation was observed in the anterior insula and pain-related brain regions to brush stimuli, thus suggesting “sumatriptan could disinhibit nociceptive signaling and make light touch less pleasant” (Kramer et al., 2007). Unmyelinated tactile afferents that are present in hairy skin only project to posterior insular cortex and serve affective aspects of tactile sensation (Liljencrantz et al., 2013). This may be of significance given the recent report that botulinum toxin inhibits mechanoreceptors by altering the neuronal surface expression of high-threshold mechanosensitive ion channels (Burstein et al., 2014).
Migraine attacks may be accompanied by sensory distortions including visual distortions (Huang et al., 2003), synesthesia (Alstadhaug and Benjaminsen, 2010), room tilt (Lopez Dominguez et al., 2007), and the ‘Alice in Wonderland Syndrome’ (Bayen et al., 2012; Ilik and Ilik, 2014). These changes may also include somatoparaphrenia (the sense of ownership towards our own body parts, that usually involve multiple brain lesions) (Gandola et al., 2012), which may be associated with “altered physiological index of perceptual analysis” to pain (Romano et al., 2014). It has been described in migraine patients that involve a diffuse number of structures (Moreira et al., 2010). In an fMRI study, abnormal, bilateral decreased activity is reported in the anterior insula in depressed patients that correlates with abnormal body perception (Wiebking et al., 2010). In patients with left posterior insular lesions, autoscopic abnormalities (visual illusory reduplication of their own body in extrapersonal space) have been described (Heydrich and Blanke, 2013).
A circuit for empathy has been defined that includes the anterior insula and anterior midcingulate cortex (Engen and Singer, 2013; Singer et al., 2009). Empathy for pain also includes the anterior insula (Singer et al., 2004). Individuals with greater empathy were found to have more gray matter in the insula which also coincided with greater functional activation within the insula region (Bernhardt et al., 2013). Intriguingly, female episodic migraine patients exhibit more gray matter in the insula than male migraineurs (Maleki et al., 2012b). While the exact functional implications are not known, it has been suggested that ‘hyperempathy’ exists in migraine patients most of whom are female (Wendt, 2010), which is thought to reflect altered interactions with the amygdala, where increased insular connectivity was observed in migraine patients (Hadjikhani et al., 2013).
Prediction error and migraine are reviewed elsewhere (Borsook et al., 2013a). Error awareness is the detection of the conscious and subconscious processing to evaluate physiological signals that are different from a baseline or homeostatic level (Borsook et al., 2013a; Ullsperger et al., 2010). The insula is considered to serve as an important hub in error awareness in a number of neurological conditions (Klein et al., 2013a).
Because the insula is involved in integrating aspects of sensory-‘ceptive’ information with emotional and motor responses (through the anterior cingulate cortex) (Nelson et al., 2010), the inability to orchestrate ‘task level control’ or attentional control may be compromised. The changes may relate to processes previously reported as “saliency, switching, attention and control” (Menon and Uddin, 2010), which may relate to the inability to exclude information in migraine (Ditchfield et al., 2006; Koppen et al., 2011; Shepherd et al., 2012; Tibber et al., 2014; Wagner et al., 2013). In figure 3F we show how altered insula function is implicated in task error through the consequential effects on the salience network (SN), a network implicated in attention and learning. In fact, errors result in activation of the SN, which is driven by the anterior right insula (Ham et al 2013 23595766).
Several approaches have been successfully used in routine clinical practice for the treatment of migraine patients, including exercise, stress reduction, pharmacotherapy, placebo, and cognitive behavioral therapy. It is conceivable that these methods mediate their effects through the insula. Below we review evidence on how treatments can affect insular cortex structure and function. It should be noted that while the insula may show activation, it might not mean that the insula mediates treatment directly, but that it may be part of an overall brain response to treatment. Treatment approaches have recently been focusing on altered interoceptive dysfunctions in other conditions such as drug addiction (Paulus et al., 2013). The use of fMRI can elucidate processes involved in functional and structural insula alterations. Figure 4 summarizes the effects on insula activation by a number of approaches used in the treatment of migraine (exercise (Fig 4A), sleep hygiene (Fig 4B), placebo response (Fig 4C)) as well as the effect of analgesic overuse as a result of treatment resulting in insula dysfunction (Fig 4D).
Exercise has been considered to help brain health through active modulation of plasticity (Cotman and Berchtold, 2002) including alterations in brain derived neurotrophic factor (BDNF) (Gomez-Pinilla et al., 2008). Aerobic exercise increases cortical gray matter in the right anterior insula (Peters et al., 2009) and inhibits gray matter loss (Gondoh et al., 2009), even in aging population (Colcombe et al., 2006). Studies have suggested that aerobic exercise improves long-term outcomes in migraineurs (Lockett and Campbell, 1992). “Training the Brain” may enhance connectivity within the insula and improve the individual’s pain and emotional processing. Indeed, diminished pain may be observed with enhanced insular health (Starr et al., 2009), though underlying processes are still not understood.
Under stressful conditions there may be increased susceptibility to migraine. The sensory insular cortex is involved in mitigating the effects of stress (Christianson et al., 2008) through safety signals (Christianson et al., 2011). Treatments that target stress may in part contribute to enhanced processing that diminishes stress through cognitive processes but also through training of the insula to limit afferent stress inputs (including pain). Indeed, there is enhanced pain in patients who anticipate pain (in migraineurs because of potentially and temporally ill defined onsets of severe headaches) because negative context may affect immediate and future pain (Phillips et al., 2003; Quartana et al., 2009). Stress may be a major contributor to allostatic load (McEwen and Gianaros, 2011) that may elicit and/or worsen migraine (Borsook et al., 2012). In fact, migraine also affects other well-known brain regions involved in the stress response. For example, decreased functional connectivity has been reported between the hippocampus and the bilateral anterior insula (Maleki et al., 2013).
The insula plays a role in interoception – a process of evaluating internal physiological states (see above). Similar to responses observed with migraine patients, dehydration also produces enhanced pain to a cold pressor test with increased activation in brain regions including the insula. With healthy dehydrated subjects, upon rehydration, these effects are diminished (Ogino et al., 2013). In the clinical setting, maintenance of good hydration has long been considered helpful for migraine patients (Ostfeld and Wolff, 1955). Given the importance of the structure in autonomic function, and in particular cardiac function, hydration may have a biological foundation in the interactions of interoceptive processes and migraine.
Migraineurs, both chronic and episodic suffer from poor sleep hygiene (Kelman and Rains, 2005; Sancisi et al., 2010; Walters et al., 2014). Figure 4B shows a reduction in activation of the right anterior insula in healthy subjects who have undergone sleep deprivation (Venkatramen 2011) further implicating the role of sleep in insula function, which may also have an effect on migraine attacks. Good sleep hygiene may therefore contribute to a balanced homeostatic and interoceptive state in migraine.
The main pharmacological treatment modalities for migraine include abortive (e.g., triptans) and preventive treatments (e.g., anti-seizure, antidepressant and potentially botulinum toxin). In a meta-analysis study of fMRI changes produced in emotional activation studies in patients with MDD, antidepressants produced a decrease in insula activation (Delaveau et al., 2011). In healthy subjects, the administration of the antidepressant escitalopram (a highly selective serotonin reuptake inhibitor (SSRI)) resulted in greater activation than overt and covert affective faces and affective words (Henry et al., 2013), suggesting that the mechanism of antidepressants in migraine patients may also modulate insular processing of emotional stimuli. As noted above, triptans produce a decrease of posterior insular cortex activation compared to the saline condition (Kramer et al., 2007). In the context of allodynia in migraine, triptans may be able to diminish the intensity and aversiveness of allodynia (Burstein et al., 2000). There is increasing excitement about the potential introduction of calcitonin gene-related peptide (CGRP) antagonists for migraine. Although most of the focus is on peripheral mechanisms, there are reports in preclinical studies of CGRP immunoreactivity in the insular cortex of rats that increase in response to aversive taste (Peyron et al., 2004). We are unaware of any imaging studies reporting insula changes as a result of specific anti-migraine therapies, whether symptomatic (e.g., nonsteroidals), abortive (e.g., triptans) or preventive (e.g., antidepressants, anti seizure) in nature. However, reports of direct pharmacological effects on insula dysfunction have been noted for paracetamol caffeine aspirin (PCA) powders in chronic migraine patients (Di et al., 2013), amitriptyline in patients with irritable bowel syndrome (IBS) and rectal pain (Morgan et al., 2005) and sumatriptan in healthy subjects (Kramer et al., 2007; Yuan et al., 2013b).
Placebo can reduce expectations (Benedetti et al., 2011), diminish pain (Colloca et al., 2013), lessen disgust (Schienle et al., 2013), and enhance treatments (Bingel et al., 2011). The opposite effect, nocebo, may enhance adverse outcomes such as pain (Colloca and Benedetti, 2007). Interestingly, two basic systems are involved – opioidergic (Petrovic et al., 2002) and non-opioidergic mechanisms (Benedetti and Amanzio, 1997). Placebo is a powerful process in migraine (Kam-Hansen et al., 2014). Accumulating data supports the notion that placebos may change the brain (Benedetti et al., 2011). Moreover, the involvement of the insula in the placebo response has been repeatedly demonstrated in imaging studies (Hashmi et al., 2012; Sarinopoulos et al., 2006; Wager et al., 2004). However, it is still unclear how best to harness the placebo effect in the clinic (Jubb and Bensing, 2013). Migraine may be a perfect disease entity to embrace and use placebo in treatments. Figure 4D shows insula activation as a result of the placebo response. Few imaging studies have been performed evaluating the insula’s role in placebo in migraine patients, but such studies should contribute to our understanding on insula function in placebo in migraine patients compared with healthy controls.
Patients with migraine, and specifically chronic migraine headaches, often pose a challenge for practitioners. Especially in patients with a higher frequency of headaches, attention to psychological and behavioral issues become a more important part of the therapeutic approach (Weeks, 2013). Cognitive behavioral therapy (CBT) may be an important and effective tool in the treatment of chronic pain and migraine, particularly when combined with pharmacological treatment. For example, a study in children and adolescents with chronic migraine showed that the use of CBT plus amitriptyline resulted in greater reductions in days with headache and migraine-related disability compared to headache education plus amitriptyline (Powers et al., 2013). CBT can be effective not only for treating chronic migraine, but also for treating comorbid diseases, such as psychiatric disorders and respiratory illness (Carod-Artal, 2014). However, it has also been suggested that individuals with (comorbid) depression and right anterior insula activation could be relatively impaired in their capacity to apply CBT-based therapies (Tegeler et al., 2014). In anxiety, successful CBT resulted in an overall down regulation of initial abnormal hyperactivity in fear-related brain regions such as the insula, amygdala, and anterior cingulate cortex (Lipka et al., 2013). Furthermore, CBT in patients with social anxiety disorder has been shown to reduce insular reactivity to fear-inducing stimuli over time (Klumpp et al., 2013). In summary, there is evidence that CBT can modulate activity in the brain that may be successful in treating migraine. The insula response as measured by brain imaging may provide an objective measure of such interventions (Klumpp et al., 2013).
The insula is involved in a wide range of functions. As such, it provides a cortical hub for the integration of extero- and interoceptive information inputs that can be transposed into higher-level behavioral function – i.e., it links “self-consciousness to the processing and integration of multisensory bodily signals” (Heydrich and Blanke, 2013). Figure 5 summarizes the interactions of putative regional functions of the insula in these functional domains as proposed for normal insula function shown in Figure 1B (Klein et al., 2013a). Accordingly, the insula constitutes an important brain region to study to further elucidate its role and plasticity in migraine and perhaps such an understanding may allow for measures that allow for therapeutic modulation of those processes that are integrated by the insula in the migraine phenotype.
This work was supported by NIH R01 (adult ####)
The authors have no conflicts of interest to report.