Frontotemporal Pathology in Motor Neuron Disease Phenotypes: Insights From Neuroimaging

Frontotemporal involvement has been extensively investigated in amyotrophic lateral sclerosis (ALS) but remains relatively poorly characterized in other motor neuron disease (MND) phenotypes such as primary lateral sclerosis (PLS), progressive muscular atrophy (PMA), spinal muscular atrophy (SMA), spinal bulbar muscular atrophy (SBMA), post poliomyelitis syndrome (PPS), and hereditary spastic paraplegia (HSP). This review focuses on insights from structural, metabolic, and functional neuroimaging studies that have advanced our understanding of extra-motor disease burden in these phenotypes. The imaging literature is limited in the majority of these conditions and frontotemporal involvement has been primarily evaluated by neuropsychology and post mortem studies. Existing imaging studies reveal that frontotemporal degeneration can be readily detected in ALS and PLS, varying degree of frontotemporal pathology may be captured in PMA, SBMA, and HSP, SMA exhibits cerebral involvement without regional predilection, and there is limited evidence for cerebral changes in PPS. Our review confirms the heterogeneity extra-motor pathology across the spectrum of MNDs and highlights the role of neuroimaging in characterizing anatomical patterns of disease burden in vivo. Despite the contribution of neuroimaging to MND research, sample size limitations, inclusion bias, attrition rates in longitudinal studies, and methodological constraints need to be carefully considered. Frontotemporal involvement is a quintessential clinical facet of MND which has important implications for screening practices, individualized management strategies, participation in clinical trials, caregiver burden, and resource allocation. The academic relevance of imaging frontotemporal pathology in MND spans from the identification of genetic variants, through the ascertainment of presymptomatic changes to the design of future epidemiology studies.


INTRODUCTION
This review explores the role of neuroimaging in characterizing frontotemporal pathology in motor neuron diseases (MNDs). While frontotemporal involvement has been extensively investigated in amyotrophic lateral sclerosis (ALS) ( Table 1), it is relatively under evaluated in other MND phenotypes, such as primary lateral sclerosis (PLS) ( Table 2), progressive muscular atrophy (PMA), spinal bulbar muscular atrophy (SBMA), spinal muscular atrophy (SMA), hereditary spastic paraplegia (HSP), poliomyelitis and post poliomyelitis syndrome (PPS) ( Table 3). Progressive muscular atrophy, SBMA, SMA, and poliomyelitis were once regarded as pure anterior horn cell disorders, but emerging data shows that the central nervous system is more widely involved than previously thought (48,50). Primary lateral sclerosis was traditionally considered a pure UMN condition, but extra-motor manifestations are now gradually recognized (29) (Figure 1). The ALS-FTD continuum of neurodegenerative disorders share common clinical, radiological, genetic, and pathological features (80,81). Similar cognitive and behavioral manifestations, however, have also been described in the non-ALS MND phenotypes (82,83). The low incidence of these slowly progressive UMN or LMN predominant disorders coupled with heterogeneous frontotemporal manifestations are all factors that may contribute to delayed or mistaken diagnoses (84)(85)(86)(87). Caregiver burden is not only heightened by diagnostic delay, but may be exacerbated by considerable behavioral challenges (88,89). Frontotemporal involvement may impact on entry into clinical trials and decision to participate in research studies, potentially leading to participation bias. From an academic viewpoint, there are synergistic efforts to evaluate frontotemporal disease burden using computational imaging in combination with clinical instruments. In parallel, these advances help to advance our understanding of disease pathology, propagation patterns, and the dynamics of anatomical spread. The objective of this review is to collate evidence from robust neuroimaging studies, distill emerging research trends, identify pertinent gaps in the literature, highlight clinical implications, and postulate research priorities in the evaluation of frontotemporal pathology across the spectrum of MND phenotypes.

METHODS
This is a focused review of original neuroimaging studies that investigated frontotemporal pathology in the following MND phenotypes; PLS, PMA, SBMA, SMA, PPS, HSP, and ALS. The search engines PubMed and Google Scholar were used to identify key papers. Individual MND phenotypes were searched paired with keywords "MRI, " "PET, " "brain imaging, " "neuroimaging, " or "frontotemporal." Only articles in English were reviewed. Editorials, opinion pieces and review articles were not selected. Additional papers were considered based on the reference list of reviewed publications. One hundred and forty-two original research neuroimaging studies were identified. Given the paucity of prospective neuroimaging studies in non-ALS MNDs, case series, neuropsychology, and post mortem studies were also reviewed in these conditions. The selected articles were systematically evaluated for cohort numbers, study design, clinical assessment, imaging methods, and anatomical focus.

Primary Lateral Sclerosis
Primary Lateral Sclerosis is an upper motor neuron disorder that typically presents with insidious spino-bulbar spasticity in adulthood (27,90). It is often associated with pseudobulbar affect that may trigger self-imposed social isolation. Extra-motor manifestations are increasingly recognized in PLS (83,91), occurring in a similar behavioral and cognitive profile to ALS (83), and rarely fulfilling the diagnostic criteria for FTD (83,91). Such deficits include impaired social cognition, executive function, verbal fluency, language, or apathy (29,82,83,91). The reported cases of frank FTD evolved several years after the insidious onset of UMN signs and were associated with progressive radiological frontotemporal atrophy (91). This is in line with the mounting body of neuroimaging evidence that supports widespread frontotemporal involvement in PLS (22,29,33).
The radiological profile of PLS varies from limited extramotor involvement to widespread pathology (27). Structural and diffusion data revealed degenerative changes in the fornix, body of the corpus callosum, anterior cingulate, dorsolateral prefrontal, insular, opercular, orbitofrontal, and temporal regions (22,27,29,33,37,42). Some studies have explored associations with underlying structural abnormalities focusing on apathy, impaired executive function, language, and verbal fluency deficits (29,33,92). Longitudinal studies have yielded inconsistent findings with regards to progressive pathology (25,26,(30)(31)(32)35). A case report described progressive cortical atrophy over an 8.5-year timeframe (93). These observations would suggest that contrary to ALS, longer follow-up intervals may be required in PLS to characterize radiological trajectories. Extramotor findings in PLS are also supported by metabolic and functional imaging studies. PET imaging studies have detected prefrontal and premotor areas of hypometabolism in PLS that are almost indistinguishable from the patterns seen in ALS (39,43,94). Whilst primarily used in a research setting, novel PET radioligand binding studies have also demonstrated alterations beyond the motor system, in the bilateral anterior cingulate gyri, and in left superior temporal lobe (2,36,39,90,95). MR Spectroscopy in PLS has mostly focused on the evaluation of the motor rather than extra-motor regions (24,35). Similar to ALS, it shows reduced N-acetyl aspartate/creatinine ratios (23,24,35,41,44) and increased myo-inositol/creatinine ratios (41) in the motor cortex suggestive of neuronal dysfunction and gliosis, respectively. Resting-state fMRI studies report increased functional connectivity in frontotemporal networks (34,92) which has been associated with executive dysfunction (4,92). Similar to ALS, increased functional connectivity is typically considered a "compensatory response" to structural degeneration (92).

Progressive Muscular Atrophy
Progressive Muscular Atrophy is a clinical diagnosis that is defined by a gradually progressive isolated lower motor neuron  disorder, evolving over many years (104). Reports of extra-motor involvement are inconsistent which is further complicated by the debate on whether PMA is a distinct entity or embedded within the spectrum of ALS (105)(106)(107)(108). There are undeniably shared clinical, radiological, and pathological features, albeit less severe compared to ALS. While the initial exclusive LMN clinical presentation distinguishes PMA from ALS, patients with PMA often later develop UMN signs (107). The cognitive profile is also strikingly similar to ALS, with varying levels of executive function, language, fluency, and memory affected (83,109). In contrast, minimal behavioral impairment is observed, and very few patients with PMA fulfill the diagnostic criteria for FTD (83,109). Some imaging studies have identified radiological abnormalities in a distribution that may explain these cognitive deficits (1,109). Structural analyses have reported loss of white matter integrity in inferior frontal, dorsolateral pre-frontal, and hippocampal regions (42). A task-based fMRI study utilizing a letter fluency task as a test of executive function showed impaired letter fluency and abnormal pre-frontal activation (47). As a counter-argument, a recent study in PMA reported preserved structural integrity with no functional connectivity alterations (4). Neither MRI spectroscopy nor PET imaging studies have identified radiological abnormalities in extra-motor regions (45,46). It is noteworthy that a dedicated neuropsychological study failed to find a difference between patients with PMA compared to controls (110). Potential shortcomings of the study designs must be considered, including small numbers of patients and the lack of sensitivity of either the chosen task or the imaging modality (46).
The shared neuropathological hallmarks also lend support to the opinion that PMA is part of the ALS clinicopathological continuum (105)(106)(107). The pathological substrates of TDP-43 positive inclusions and occasional fused-in-sarcoma (FUS)positive basophilic inclusions are observed in both conditions, but at a lesser burden and more limited distribution in PMA (108,111). Post-mortem studies in PMA typically describe LMN degeneration, occasional pyramidal tract degeneration, and additional TDP-43 positive inclusions in the primary motor cortex and hippocampus even in the absence of UMN degeneration (108,111). These findings raise the question, if in fact the results of PMA studies should be streamlined, interpreted, and analyzed under the umbrella of ALS.

Spinal Muscular Atrophy
Spinal Muscular Atrophy is an autosomal recessive disorder that is caused by either homozygous deletions or loss of function mutations in the survival motor neuron 1 (SMN1) gene resulting in a deficiency of survival motor neuron (SMN) protein (112). It typically manifests as a proximal, predominantly symmetrical motor weakness. The phenotype is stratified in levels of decreasing severity from type 0 to type IV, depending on age of symptom onset and achievement of developmental milestones (112). There are preliminary signals of cerebral involvement in the more severe phenotypes, but it is not yet clear if there is preferential involvement of frontotemporal regions.       The only two cross-sectional quantitative multimodal MRI brain studies evaluated the same 25 treatment naïve adults with type III or type IV SMA initially focusing on the cerebrum and then the cerebellum (48,50). No supratentorial cortical atrophy was detected (50), but focal cerebellar changes were noted. In the more severe clinical phenotypes, qualitative MRI brain scans have captured more dramatic findings (49,(113)(114)(115). In type 0 SMA, widespread supratentorial, and sometimes infratentorial, brain atrophy has been reported. A longitudinal case series of patients with type 0 SMA showed interval radiological abnormalities involving the thalamus and basal ganglia (49,114). Similar radiological findings have been described in type I SMA (115). For the most part, neuropsychological studies demonstrate preserved cognition (116)(117)(118)(119). This is with the caveat that these studies are mostly limited to children and omit the more severe clinical phenotypes (120). Some aspects of childhood development are even deemed superior compared to healthy controls (121)(122)(123)(124)(125). The only neuropsychological study of adults with type II or type III SMA described normal rather than superior cognitive abilities. This study reported a possible adaptive mechanism of an inverse correlation between executive function and physical ability, but the level of executive function did not exceed healthy controls (116). In contrast, there are indications of attention and executive function deficits in children with type I SMA (120,126).
The post mortem examination of the brain is often confounded by coexistent anoxic changes (127,128). The more severe clinical phenotypes display more widespread features of degeneration involving the cerebral cortex, thalamus, brainstem, and some cranial nerve nuclei that are congruent with ante mortem radiological abnormalities (49,129). Most of these regions seem spared in the milder phenotypes (127,128,130). This has been interpreted as selective neuronal network degeneration occurring below a threshold of SMN protein, although the true clinical significance of this is unknown (129). Overall, the radiological characterization of the more severe clinical phenotypes has proven challenging because of the rarity of the condition, significant disability, and limited life expectancy. In the advent of gene therapy, there may be opportunities for future research in this cohort.

Spinal and Bulbar Muscular Atrophy (Kennedy's Disease)
Spinal and Bulbar Muscular Atrophy, also known as Kennedy's disease, is an X-linked trinucleotide repeat disorder due to expansion of cytosine-adenine-guanine (CAG) repeat in the androgen receptor gene (131). It is a multisystem disorder that typically presents in men in their fourth decade of life with slowly progressive weakness, bulbar involvement, and muscle atrophy due to insidious lower motor neuron degeneration (132,133). Relatively mild cognitive deficits have been consistently described (134)(135)(136). While it is a multi-system disorder, the involvement of the central nervous system has been relatively under-evaluated from a radiological viewpoint (131).
The few brain imaging studies indicate various degree of frontotemporal involvement (53,54,56). Quantitative MRI analyses demonstrate a spectrum of frontal gray and white matter abnormalities ranging from entirely unaffected to subtle gray matter atrophy and extensive white matter degeneration (53,56). Widespread loss of white matter integrity has been reported in the brainstem, corticospinal tracts, and limbic system (40,51,56). A single PET imaging study showed hypometabolism in frontal areas (54). The results of conflicting MR spectroscopy studies highlight that subclinical neuronal dysfunction may not be detected by certain imaging protocols (52,55). A long echotime MR spectroscopy study demonstrated altered metabolite ratios in the brainstem and motor regions (52); however, a short echo-time MR spectroscopy study failed to reproduce these findings (55). The discrepancy in these results may be explained by the potential pitfall of artificial metabolite elevation because of either metabolite signal overlap or incorrect baseline determination in short echo-time MR spectroscopy (137). These radiological findings are complemented by consistent reports of neuropsychological dysfunction in this cohort albeit mostly at a subclinical level (134)(135)(136). Deficits may be so subtle that performance on standard tests of executive function can be normal (134,135). Mild deficits in social cognition have also been recorded (135). In contrast, single cases of more severe frontal dysfunction have been repeatedly described (138,139).
Most post-mortem studies in SBMA focus on cardinal spinal cord, peripheral nerve, and proximal muscle changes (133,139,140). The pathological examination of cerebral hemispheres is seldom reported. A post-mortem report of an SBMA patient with significant cognitive impairment demonstrated marked diffuse subcortical gliosis in the pre-frontal region, hippocampus, and the degeneration of fronto-bulbar fibers in the midbrain without accompanying cortical pathology (139). Immunohistochemical studies have shown that the pathogenic nuclear mutant AR protein is present in abundance in the central nervous system; supporting the rationale to systematically evaluate cerebral changes in future SBMA studies (141).

Poliomyelitis and Post-polio Syndrome
Post-polio syndrome is characterized by progressive muscular weakness with or without pain, fatigue, and muscle atrophy in patients who have recovered from a distant polio infection (61). Patients often report diverse cognitive symptoms, mostly deficits in attention or memory; however objective evidence is strongly confounded by comorbid factors such as fatigue (57,(142)(143)(144). The reportedly high prevalence of extra-motor symptoms is contrasted by the relative lack of cerebral radiological abnormalities in post-polio syndrome (59).
A quantitative MRI study detected minimal cortical and subcortical atrophy, involving the cingulate gyrus, temporal pole, and left nucleus accumbens (59). These subtle changes were not appreciated in other studies (60,61). Qualitative MRI studies either identified no abnormalities or discrete subcortical hyperintensities that were hypothesized to contribute to the disabling comorbid fatigue (57,58,145). Patients with postpolio syndrome frequently exhibit high levels of self-reported fatigue, apathy, and verbal fluency deficits. In the absence of widespread frontotemporal imaging abnormalities, these extramotor symptoms are postulated to be multifactorial in origin with factors such as low mood, poor sleep, and polypharmacy all playing an additive role (59).
These observations are corroborated by historical pathological studies that demonstrate preferential involvement of the brainstem rather than the cerebrum (57,146,147). This is further complicated by reports of patients with a history of polio, who later develop sporadic ALS and demonstrate mixed neuropathological features including the hallmarks of both diseases (148,149).

Hereditary Spastic Paraplegia
Hereditary spastic paraplegias (HSPs) are a clinically and genetically heterogenous group of neurodegenerative disorders that present as progressive limb weakness and spasticity. They were traditionally divided into "pure" or "complicated" phenotypes based on the absence or presence of extra-motor involvement, respectively (63). In recent times, there has been a shift to stratify these cohorts in accordance with their genetic diagnoses (150). Interestingly, there are radiological indicators of frontotemporal dysfunction irrespective of the subgroup.
Brain imaging studies have shown a reduction in whole brain volume in both clinical phenotypes (68). In pure HSP the volume of gray matter volume is thought to be mostly preserved, whereas in complicated HSP the volume of cortical and deep gray matter may be reduced. The only longitudinal study detected no change in cerebral volume over a 2-year followup period (72). This is with the caveat that longer time intervals may be required to detect a significant change. Loss of white matter integrity has been identified in the corpus callosum, in the frontotemporal and parietal regions in both groups (40, 62-64, 72, 73). The severity of these findings correlate with the degree of cognitive impairment (63). Given the relative cortical sparing, cognitive deficits in these cohorts were postulated to be primarily subcortical in origin (64,66,67). This was supported by MR spectroscopy reports of abnormal metabolic ratios in the subcortical white matter (65,72,77,(151)(152)(153)(154)(155). PET imaging studies detected cortical hypometabolism, sometimes implicating the frontotemporal regions (76,(156)(157)(158)(159)(160)(161). This was accompanied by clinical measures of frontal dysfunction (76,161). Restingstate fMRI studies have shown altered functional connectivity involving the primary motor cortex, insula, and superior frontal gyrus (70). Task-based fMRI studies typically report abnormal activation patterns in sensorimotor areas whilst performing motor tasks (69,78).
In the advent of genotyping, there has been a focused effort to define the radiological signatures of specific genotypes. Spastic paraplegia 4 (SPG 4) is the most common autosomal dominant HSP subtype that is characterized by widespread white matter degeneration with relatively preserved gray matter (64,71,75). Subclinical cognitive deficits have been described that later follow a more rapid trajectory of decline escalating in the eights decade of life (162)(163)(164)(165). Spastic paraplegia 11 (SPG11) and spastic paraplegia 7 (SPG7) are rare autosomal recessive HSP subtypes that reveal white matter degeneration involving the frontotemporal regions amongst other features (67,74,79,150,152,(166)(167)(168). Varying degrees of cognitive deficits including attention, memory, and executive dysfunction have been described in these genotypes and others (66,79,169).
The few post-mortem studies corroborate the radiological descriptions of frontotemporal pathology. Autopsy reports of those with a clinical rather than genetic diagnosis must be interpreted with caution. In clinically defined cases, marked cerebral atrophy and severe gliosis of the cerebral white matter has been described sometimes preferentially involving prefrontal and frontal areas (170)(171)(172). In SPG11, widespread frontotemporal cortical degeneration has been described (85). Similar pathological observations have been reported in SPG4, in addition to widespread ubiquitin positivity (164).

Amyotrophic Lateral Sclerosis
Amyotrophic lateral sclerosis is the most common form of MND that is characterized by progressive upper and lower motor neuron degeneration in the motor cortex, brainstem nuclei, and anterior horn of the spinal cord. It begins with progressive limbonset or bulbar-onset muscle weakness that clinically manifests as cramps, fasciculations, muscle wasting, difficulty swallowing, or speaking before ultimately advancing to respiratory failure (173). Additional cognitive and/or behavioral impairment is universally recognized and a minority of patients with ALS also fulfill the diagnostic criteria for FTD (174).

DISCUSSION
This review collates evidence of radiological frontotemporal involvement in common MND phenotypes. Existing neuroimaging studies suggest that frontotemporal degeneration may be readily detected in ALS and PLS; a varying degree of frontotemporal pathology may be captured in PMA, SBMA, and HSP. Cerebral involvement without regional predilection may be exhibited in the more severe clinical phenotypes of SMA; and there is limited evidence for cerebral changes in PPS (Figure 2, Table 4). These radiological features may precede clinical symptoms, and longitudinal studies often capture gradual progression. Imaging studies in MND suffer from considerable inclusion bias because of disease-specific factors. Patients with significant apathy, motor disability, respiratory compromise, or sialorrhoea are less likely to participate or return for follow-up imaging.
In clinical practice, the wide spectrum of frontotemporal manifestations in ALS are already incorporated in the clinical diagnostic criteria (174). It is anticipated that these features will be a fundamental part of future revisions, in conjunction with supportive neuroimaging data (258). Despite implications for survival (259), clinical staging systems of ALS have omitted to include a cognitive facet thus far (260,261). These observations have also not yet translated into the diagnostic criteria of other MND phenotypes. While such deficits are increasingly recognized in PLS, they are deemed too infrequent to be included in the core clinical features (262). The link between FTD and other rare MND phenotypes may have important implications for everyday clinical practice, particularly given that many non-ALS MND phenotypes are associated with longer survival than ALS (263). The awareness of possible frontotemporal dysfunction may prompt the use of neuropsychological screening tests in the routine evaluation of these patients. Validated, disease-specific screening tools are preferred to generic instruments, and these are available in ALS (264,265). Several of these are adapted to motor disability and dysarthria, and interrogate domains commonly affected in ALS. It is worth noting that patients with predominant frontotemporal cognitive deficits should be screened for incipient motor deficits (266). The early recognition of neuropsychological deficits is crucial for individualized patient care including: the appraisal of decision-making capacity, caregiver support, resource allocation, and the anticipation of management challenges (89). It may also allow clinicians to consider pharmacological and non-pharmacological interventions such as cognitive or behavioral rehabilitation. In the context of FTD, this is primarily focused on developing compensatory skills for adapting to functional impairments with the lowest level of assistance required. For example, an electronic device calendar is a daily planning tool that can be used to establish routines and set reminders to initiate activities such as taking medications (267). There is also an evolving interest in early language therapy interventions (268,269). The education of caregivers is crucial to identify unmet needs of the patient that may trigger behavioral problems. These measures have proven to be beneficial to both the patient and their caregivers (267).
In tandem, technological innovations have enriched the supportive radiological data. High-field MRI generates better quality images and acceleration techniques enable shorter data acquisition that may be better tolerated by patients. Quantitative MRI analyses using validated computational pipelines and reliance on robust comparative, correlative, and classifier models enhance the clinical interpretation of vast imaging datasets (270). The advent of structural and functional connectivity studies have ignited interest in the concept of disease-specific selective network degeneration rather than the emphasis on focal pathology (271). These methods have proven particularly useful to differentiate clinical phenotypes and map longitudinal changes in neurodegenerative disorders (7,271). Novel MRI pulse sequences, non-Gaussian diffusion models such as DKI or NODDI, quantitative susceptibility mapping, and multi-voxel spectroscopy are just some of the promising new tools enriching our armamentarium of imaging tools (272)(273)(274). While these methods continue to be tested in the research community they have not been implemented in routine clinical radiology protocols (275).
Frontotemporal involvement across the spectrum of MND phenotypes has important implications for clinical trials. It invites the opportunity for the development of radiological biomarkers that quantify and track frontotemporal involvement (276). Pioneering gene therapy trials have primarily focused on clinical outcome measures such as motor milestones, requiring artificial ventilation, and survival (277)(278)(279). This is also relevant because there has been interest in developing adjunctive interventions such as transcranial or neuromuscular magnetic stimulation (280,281). These methods are not only applicable to symptomatic patients but also to those in the pre-symptomatic stages of their disease. In ALS, genotype-specific radiological alterations have been detected in pre-symptomatic carriers of pathogenic mutations decades before the onset of clinical symptoms (282)(283)(284). Awareness of associated behavioral and cognitive impairment allows for due preparation and adaption of study designs if required (285).
As we begin to incorporate these developments into clinical practice and clinical trial designs, there are pressing academic questions to be elucidated. First and foremost, it is uncertain if the motor or extra-motor changes evolve in sequence or in parallel across all phenotypes. This topic is probably best explored in ALS where there is a unique opportunity to study the presymptomatic phase in carriers of pathogenic C9orf72 repeat expansions. In this cohort, radiological co-existence of motor and extra-motor involvement has been consistently described (282)(283)(284)(285)(286). Overall, the topography of radiological alterations is largely similar but less marked than what is described in symptomatic cases. It is unclear if the initial pattern dictates the ultimate clinical phenotype given that both FTD and ALS may have co-existent subclinical motor and extra-motor manifestations (284,287,288). It is also unclear whether these findings solely represent early neurodegenerative changes; some postulate that they capture a developmental abnormality (284). From a clinical perspective, early cognitive deficits have been described in presymptomatic carriers of C9orf72 expansion before the phenotype is defined (282)(283)(284)286). The notion of cognitive reserve has been increasingly evaluated in ALS which may impact on the sequence of symptom manifestation. It suggests that those with a high level of cognitive reserve, often proxied with educational attainment, require a greater degree of brain pathology to meet the threshold for clinical symptoms (289). This concept has been investigated in greater detail in FTD, but similar themes are also emerging in ALS. The level of cognitive reserve appears to predict cognitive performance and the degree of brain imaging abnormalities (290)(291)(292). These observations suggest that patient-specific factors influence the chronology of clinically evident symptoms. Some argue that the debate of whether extramotor or motor symptoms emerge first in ALS is antiquated and that cognition and motor function are inseparably intertwined (293). It is hypothesized that the selective deficit in action words and verb processing detected in patients with ALS is in fact a cognitive manifestation of motor dysfunction (293). Although some disagree and consider it to be a feature of executive dysfunction (294). Task-based fMRI studies in healthy controls have consistently shown that reading action words activates areas along the motor strip that were responsible for conducting these movements (295). In ALS, action observation and motor imagery are routinely utilized in fMRI studies to compensate for motor disability (296).
Cognitive deficits in specific domains have been linked to the degeneration of single structures in MND (1,14,16,33,186,188,244). Often there is frank dissociation between cognitive and radiological findings (59,63), but a reporting bias for confirmed associations prevents the gauging of this occurrence. Correlation analyses in ALS linked apathy to anterior cingulate and accumbens nucleus degeneration (193)(194)(195), and memory impairment to hippocampal degeneration (13). Linking cognitive deficits to single structures however may be a reductionist approach, which overlooks the role of complex cortico-subcortical networks in mediating cognitive functions (297). Accordingly, the underpinnings of neuropsychological deficits are probably best evaluated at a circuitry-integrity level instead of seeking associations with focal structures (271). Traditional structural measures are increasingly complemented by connectivity metrics to appraise the integrity of functional circuits (240,242). The selective vulnerability of functional networks is thought to drive cardinal manifestations of neurodegenerative conditions (271). It may or may not be associated with focal atrophy of crucial nodes within these networks (298).  There are stereotyped shortcomings in the current literature that remain to be addressed. First, the low incidence of these conditions leads to small sample size despite multi-center collaborations. Second, while case-control study designs are often used to evaluate these rare disorders, this cross-sectional approach is suboptimal to characterize dynamically evolving processes. Furthermore, the indolent progression of the non-ALS MND phenotypes may require relatively long follow-up intervals to detect progressive radiological changes (72). Third, co-existing neurodegenerative disorders are potential confounders, such as behavioral variant Alzheimer's dementia. To account for this, the use of serum or cerebrospinal fluid biomarkers should be considered in future study methods to enhance diagnostic certainty. Fourth, there is a scarcity of pre-symptomatic studies and often these cohorts are not followed longitudinally until phenoconversion. Fifth, the diagnostic criteria are not welldefined in some MNDs (104). The diagnosis of "definite PLS" requires a symptom duration of at least 4 years which may further limit the number of patients available for recruitment (28). Sixth, imaging studies often concentrate on supratentorial cortical regions, overlooking the contribution of subcortical and cerebellar pathology to cognitive and behavioral manifestations. The sensitivity limitations of single imaging modalities are seldom acknowledged. Subtle abnormalities may not be detected, considerable neuronal loss may ensue before it becomes radiologically evident. Seventh, the practical implications of cognitive deficits need to be specifically investigated. The presence of cognitive impairment in ALS is considered a negative prognostic indicator that is associated with increased caregiver burden, reduced quality of life, and reduced survival; whereas the implications of cognitive impairment in other MND phenotypes is woefully under-evaluated despite their markedly longer survival (259). Finally, there is a disappointing lack of post-mortem validation of radiological findings. This is further complicated by the inherent bias of the pathological literature to favor atypical cases that are unlikely to represent the true hallmarks of these conditions. This paper offers an overview of imaging efforts across the spectrum of MNDs to investigate frontotemporal disease expansion. It highlights the disproportionate emphasis on ALS, which offers valuable lessons to conduct similar studies in other MND phenotypes. Radiological observations highlight the rationale for routine screening for frontotemporal dysfunction to inform individualized patient care. The quality of the data may be enhanced by using multiparametric imaging protocols, longitudinal study designs, and the inclusion of pre-symptomatic cohorts where possible. The opportunity for international collaborations through carefully harmonized protocols should be explored to maximize the number of study participants in low incidence phenotypes.

CONCLUSIONS
In contrast to ALS, the quantitative characterization of frontotemporal disease burden in non-ALS MND phenotypes remains under investigated. The nuanced evaluation of frontotemporal dysfunction across the entire spectrum of MNDs has important pragmatic implications for individualized clinical care, caregiver support, clinical trial designs, and more broadly, for our understanding of disease biology.

AUTHOR CONTRIBUTIONS
Manuscript drafted by MM and PB. Review for intellectual content by PCor, PCou, WS, and P-FP. All authors contributed to the article and approved the submitted version.