Understanding Cerebellar Contributions to Cognitive and Affective Functions in Aging and AD/ADRD

Agenda

See below for agenda information for each day.

Day 1 | Sept. 12, 2023

Welcome and Keynote Address

10:00 a.m. Welcoming Remarks, Eliezer Masliah, M.D., NIA, NIH

10:10 a.m. The Cortico-Cerebellar System: Organisation and Function, Narender Ramnani, Ph.D., Royal Holloway University of London

10:40 a.m. Keynote Q&A

• Moderator: Matt Sutterer

Session 1 | Cerebellar Contributions to Cognitive Aging

10:50 a.m. On Scaffolding and Sex: The Cerebellum in Females Across Adulthood and Aging, Jessica Bernard, Ph.D., Texas A&M, Session Chair

11:10 a.m. Theory and Neuroscience of Cerebellar Cognition and Emotion: Clinical Insights and Implications for Neuropsychiatry Across the Age Spectrum, Jeremy Schmahmann, M.D., Harvard Medical School

11:30 a.m. Break

11:40 a.m. Resilience and Vulnerability: Neuropathological Measures of Human Cerebellum Morphology and Circuits in Aging and Dementia, Erik Carlson, M.D., Ph.D., University of Washington

12:00 p.m. The Cerebellum as a Potential Treatment Target for Late-life Cognitive and Mood Disorders, Vonetta Dotson, Ph.D., Georgia State University

12:20 p.m. Session 1 Panel Discussion and Q&A

• Moderator: Jessica Bernard

1:05 p.m. Lunch Break

Session 2 | Preclinical Evidence of Non-motor Cerebellum Functions

2:05 p.m. The Cerebellum as a Context Machine for Action and Thought, Sam Wang, Ph.D., Princeton University, Session Chair

2:25 p.m. Social Memory Deficit Caused by Dysregulation of the Cerebellar Vermis, Yi-Mei Amy Yang, Ph.D., University of Minnesota

2:45 p.m. Reward Signals in Cerebellar Circuits, Michael Hausser, Ph.D. International Brain Laboratory

3:05 p.m. Cerebellar Modulation of the Brain's Dopaminergic Circuitry, Kamran Khodakhah, Ph.D., Albert Einstein College of Medicine

3:25 p.m. Session 2 Panel Discussion and Q&A

• Moderator: Sam Wang

4:10 p.m. Adjourn for Day

Day 2 | Sept. 13, 2023

8:45 a.m. Welcome and Recap of Day 1, Matt Sutterer, Ph.D. and Coryse St. Hillaire-Clarke, Ph.D.

Session 3 | Cerebellar Roles in Affective Processing, Social Function, and Cognition

8:55 a.m. Cerebellar Circuits, Timing, and Cognition, Krystal Parker, Ph.D. University of Iowa, Session Chair

9:15 a.m. Computational Modeling of Cerebellar Stimulation Using Lobule-Specific Dosage and MRI-Based Approach, Zeynab Rezaee, Ph.D., NIH/NIMH

9:35 a.m. Specifying Constraints on Cerebellar Contributions to Prediction, Richard Ivry, Ph.D., University of California Berkeley

9:55 a.m. The Human Cerebellum in Social and Emotional Processing, Ingrid Olson, Ph.D., Temple University

10:15 a.m. Session 3 Panel Discussion and Q&A

• Moderator: Krystal Parker

11:00 a.m. Break

11:15 a.m. Overall Discussion

• Moderator: Coryse St. Hillaire-Clarke

12:00 p.m. Closing Remarks, Matt Sutterer, Ph.D.

12:05 p.m. Adjourn

Executive Summary

The Understanding Cerebellar Contributions to Cognitive and Affective Functions in Aging and AD/ADRD was held on Sept. 12-13, 2023. This summary highlights findings and conclusions for each of the discussions.

Executive Summary

While the cerebellum has traditionally been associated with motor functions, a growing body of work has identified a contributing role for the cerebellum in several cognitive and affective functions. Despite known changes in aging with respect to both cognition and motor performance, most aging work up to this point has focused on associating these performance changes to neural changes in the cerebral cortex. On September 12 and 13, 2023, the National Institute on Aging (NIA) convened a panel of experts in cerebellar research to explore the cerebellum’s contributions to cognitive aging, the observed neurobiology and symptoms of Alzheimer's disease (AD) and Alzheimer's disease-related dementias (ADRD), as well as its potential role in the development and maintenance of cognitive reserve.

The meeting opened with a keynote address on the organization and function of the cerebello-cortical system. Cerebellum-centered forward models of executive control describe cerebellar regulation of motor and cognitive behavior, in which a copy of initial cortical commands is represented in functionally related cerebellar regions. Research using neuronal tracing methods to study cerebro-cerebellar connections in animal models reveals cerebellar linkage with multiple sensorimotor, associative and paralimbic cortical regions, including the orbitofrontal and anterior cingulate cortices. The translation of cerebro-cerebellar connectivity studies from monkey to human using blood-oxygen-level-dependent (BOLD) functional magnetic resonance imaging (fMRI) was found to have limitations when inferring age-related differences in neural activity, because the amplitude of the BOLD response decreases as people age, while signal dispersion and peak latency increase. These results suggest that canonical hemodynamic response functions (HRF) are not universally applicable for analyzing subject groups across the lifespan. Further, the canonical HRF showed higher within-group BOLD signal variance for younger subjects, leading to potential false negative results.

Cerebellar Contributions to Cognitive Aging

The cerebellum receives feedforward projection fibers by way of the basis pontis from most areas of the cerebral cortex, arranged with a high degree of topographic precision. The consistency of cerebellar architecture has led to the concept of a universal cerebellar transform of information processing, which is applied to the diverse streams of cerebro-cerebellar connections thus enabling behavior. Several studies suggest that the cerebellum contributes to motor, memory, and cognitive functioning through scaffolding, in which the cortex offloads processes to the cerebellum through efferent signal copies. Findings from these studies suggest that (1) older adults may be less able to offload cortical processes to the cerebellum due to degraded connections, and (2) intact scaffolding may be an indicator of aging resilience.

Several studies have demonstrated sex-related differences in cerebellar structure in older adulthood. Compared to males, females experience more negative age-related impacts on global cognitive scores and have a higher estimated lifetime risk of AD. Furthermore, hormonal transition periods (e.g., puberty, menopause) are associated with changes in grey matter, and cerebellar functional connectivity patterns vary with hormone levels across the menstrual cycle. Several studies have evaluated cerebellar contributions to the development of AD symptoms in addition to amyloid plaque burden, which have been observed in the cerebellum during late stages of AD. These studies found that dementia patients have lower cerebellar volume, reduced granular layer neuropil, and loss of excitatory synapses in mossy fiber rosettes (i.e., the sole input from ponto-cerebellar communication). These changes are associated with multiple dementia-related measures, such as dementia status, Braak and Thal stage, Consortium to Establish a Registry for Alzheimer’s Disease (CERAD) score, age of death, and posterior fossa weight.

The investigation of cerebellar lesions has been informative for understanding connections between the cerebral cortex and the cerebellum. For example, lesion location in cerebellar stroke patients can determine whether the patient exhibits motor or cognitive deficits: lesions in the anterior lobe result in cerebellar motor syndrome, whereas lesions in the posterior lobe primarily lead to symptoms of cerebellar cognitive affective syndrome, a syndrome defined by dysregulation of executive functions, spatial cognition, language deficits, and personality change.

Preclinical Evidence of Non-Motor Cerebellar Functions

Animal studies can advance understanding of cerebellar signaling pathways by revealing circuit- and cellular-level functions, including (1) the directionality of signaling effects through viral tracing techniques, (2) circuit- and cell-type specificity, and (3) perturbational effects on neural activity, tasks, and behavior that may be beneficial for recovery of cerebellar function. Regarding the first, studies using viral vector tracing techniques helped produce a map of amygdalar integration in the cerebello-cortical circuits involved in cerebellar vermis-dependent social recognition memory. Next, findings from several studies using light-sheet microscopy of cerebellar c-Fos suggest that lobule VI plays a role in thalamic and brain-wide coordination of learning behaviors. Finally, optogenetic techniques revealed (1) that Crus I may enhance the salience of sensory stimuli for working memory tasks, and (2) that the cerebellar vermis plays a role in the retrieval of social memory, but not its encoding. In vivo techniques like these can support research on the cerebellum’s role in the development of autism spectrum disorder (ASD) and in teaching priors during brain maturation.

Animal studies have also advanced research on reward signals in cerebellar circuits. Mice trained to move an on-screen object to a target demonstrated automation of motor behavior with increased task experience. Purkinje cells (PCs) rarely expressed the error signals consistent with forward model activation theories. Instead, the team found robust signals associated with the reward administered during the tasks. Further clustering analyses of activated neurons identified functionally-defined PC populations in regionally distinct microzones which could be further categorized as either reward-activated or reward-suppressed. These findings support previous studies of reward signaling in the cerebellum and resemble dopaminergic signaling in the ventral tegmental area (VTA), which receives projections from the cerebellum. Optogenetic activation of the cerebellar-VTA pathway rapidly increased synaptic firing in the medial prefrontal cortex (mPFC) as well as nucleus accumbens (NA) activation, suggesting that the cerebellum can modulate the VTA and its dopaminergic targets. These support a model in which the cerebellum performs similar roles in both cognitive and motor circuits.

Cerebellar Roles in Affective Processing, Social Function, and Cognition

The uniform cerebellar transform may represent a neural computation of prediction-based modeling that can apply to both motor and cognitive tasks, with notable constraints on scope. This transform may lie on a continuum from cerebellum-dependent to cerebellum-independent predictions. The homogenous architecture of the cerebellum supports the theory of the uniform cerebellar transform by providing a feasible organization for such a model. The theory is further supported by evidence from motor, visual cognition, and arithmetic tasks that suggest transformation across motor and non-motor domains. However, this theory may not hold for language-based tasks, suggesting that language may be mediated by memory systems instead.

Connections between the cerebellar dentate nuclei and prefrontal cortex could inform cerebellum-targeting therapeutics for cognitive dysfunction, especially for behavioral processes that typically decline during aging or neurodegenerative disease progression. Low-frequency stimulation therapy studies rescued timing behavior in patients with mood disorders, suggesting that the cerebellum mediates low-frequency oscillation rhythms regulated by dopaminergic signaling. To characterize the utility of these low-frequency rhythms as clinical treatments, ongoing investigations are examining the effects of cerebellar neurostimulation therapies in patients with mood disorders and other neurological and psychiatric conditions.

The unique geometry of the cerebellum poses challenges in non-invasive stimulation treatment efficacy. Age-related increases in cerebrospinal fluid (CSF) volume and reductions in cerebellar volume reduce magnitude and directionality of stimulation frequencies, indicating that therapeutic stimulation might have reduced efficacy for older adults who already experience cerebellar deficits. Computational models and experimental validation may improve treatment specificity and reduce variability.

The cerebellum may also play a mediating role in affective behaviors and mood, as evidenced by the effects of cerebellar insult on social behavior. Although children who sustain cerebellar injury show high rates of altered social behavior, the impact of cerebellar insult in older adults is less consistent. The cerebellar regions activated during social behaviors project to mood-associated cortical regions, with the increased density of these connections predicting increases in anxiety and depression in older adults.

Overall Meeting Discussion

The meeting concluded with an overarching discussion of the limitations of fMRI, BOLD, and other cerebellar imaging techniques, as well as a three-step process to address how the cerebellum engages with neuromodulatory systems in the brain: (1) mapping connectivity of the cerebellum in humans; (2) studying how connectivity changes in the context of aging and neurodegenerative disease; and (3) using animal models to test genetic manipulation treatments on target cerebellar regions. Participants concluded that conducting secondary analyses on large repositories of human cerebellum imaging could help correlate neurophysiological studies of animal models with functional neuroimaging data from humans. In addition, they agreed that researchers should consider beginning cerebellum physiology studies in aged wildtype rodents before identifying or developing transgenic models.

Meeting Summary | Day 1

Welcome Remarks

Eliezer Masliah, M.D., NIH/NIA

Dr. Eliezer Masliah welcomed the meeting participants and outlined the meeting goals. Although the cerebellum has traditionally been associated with motor functions, a growing body of research has identified a contributing role for the cerebellum in several cognitive and affective functions. Despite known changes in aging associated with both cognition and motor performance, most aging research has focused on changes in the cerebral cortex. Thus, this meeting aimed to explore the cerebellum’s role in the development and maintenance of cognitive reserve, as well as the cerebellum’s contributions to the observed neurobiology of aging, Alzheimer's disease (AD), and Alzheimer's disease-related dementias (ADRD).

Keynote Address—The Cortico-Cerebellar System: Organization and Function

Narender Ramnani, Ph.D., Royal Holloway University of London

Dr. Narender Ramnani’s lab characterizes motor and cognitive behavior in cerebellum-centered forward models of executive control. When motor commands are executed, a copy of the initial command traveling through the motor cortex is represented in associated cerebellar regions. The cerebello-cortical connectivity mediating this projection is thought to support the learned automation of both motor and cognitive operations. Dr. Ramnani’s team used the Cambridge Center for Ageing and Neuroscience (Cam-CAN) dataset to assess the impacts of aging on these cerebello-cortical connections and to demonstrate the limitations of using hemodynamic response functions (HRF) to analyze lifespan-oriented datasets.

Conventional neuronal tracing methods have revealed motor projections from the cortex to the pontine nuclei and cerebellum. Dr. Ramnani’s lab selected 59 contiguous areas of the frontal lobe correlated with motor execution to investigate the topographical organization of cortical connections to the cerebellum. Cerebello-cortical connectivity increased from the anterior to posterior frontal lobe, with the strongest concentration of neural projections in the middle frontal gyrus. Non-motor associated regions such as the orbitofrontal and anterior cingulate cortices also demonstrated cerebellar connectivity. The team found no age-related increases in cerebellar connectivity in any of these regions, whereas they found age-related reductions in 44 of the 59 areas. These reductions were concentrated in areas in and around the precentral gyrus, predominantly affecting vermal regions of the fifth and sixth cerebellar lobules.

A second study investigated the limitations of using blood-oxygen-level-dependent (BOLD) functional magnetic resonance imaging (fMRI) to infer age-related differences in neural activity. The standard BOLD response measurement uses generalized linear models constructed from regressors to create a canonical HRF model with a fixed time course. However, if aging affects neuronal activity and vasculature, it may also affect the time course of the BOLD response. Dr. Ramnani’s team found that the amplitude of the BOLD response decreases as people age, while signal dispersion and peak latency increase. These results suggest that canonical HRF is not universally applicable for analyzing subject groups across the lifespan. Further, the team observed that the canonical HRF contributes to higher within-group BOLD signal variance in younger subjects. Therefore, fMRI studies focused on young populations could yield more false negative results. However, using Fourier basis sets introduced flexibility to the canonical HRF model and decreased the likelihood of error.

Keynote Q&A

Moderator: Matt Sutterer, Ph.D., NIA

Participants asked how Dr. Ramnani’s BOLD findings might account for time lag in cerebellar resting state electroencephalography (EEG). Dr. Ramnani noted that the team had not directly studied this question but observed that generalized linear models should already adjust any cerebellar imaging variability because BOLD signals are measured within a fixed time course. Participants also asked how to account for age-related increases in within-group variability for connectivity data. Although Dr. Ramnani’s lab has not observed any such increases, generalized linear models can also resolve this variability. Dr. Ramnani also discussed whether technical issues could explain some of the lab’s findings in their analysis of connectivity declines, emphasizing the need for care when scaling up small effect sizes or ascribing projections to specific brain regions. When asked if the team observed any connections preserved during aging, Dr. Ramnani highlighted the 15 areas of the frontal lobe that showed no cerebellar connectivity changes. Dr. Ramnani added that much of the cerebellar cortex has yet to be explored and might contribute insights on cerebellar connectivity to non-motor associated regions. The dramatic declines in cerebello-cortical connectivity that Dr. Ramnani’s team observed are comparable to similar declines in medial portions of the interior cortex. Although the team did not detect interaction effects between age and physical activity, Dr. Ramnani acknowledged that measuring individual reaction times would be important to attribute connectivity changes.

Session 1: Cerebellar Contributions to Cognitive Aging

Session Chair: Jessica Bernard, Ph.D., Texas A&M University

On Scaffolding and Sex: The Cerebellum in Females Across Adulthood and Aging

Jessica Bernard, Ph.D., Texas A&M University

Several studies suggest that the cerebellum contributes to motor, memory, and cognitive functioning through scaffolding, in which the cortex offloads processes to the cerebellum through efferent signal copies. Dr. Jessica Bernard’s group conducted several cross-sectional studies demonstrating that older adults and younger adults exhibit differences in lobular structure, particularly in Crus I and II and lobules IV and V. Next, the group found older adults showed less connectivity from the cerebellum to the lateral prefrontal cortex, medial temporal lobes, and striatal regions. The group also evaluated functional performance, finding that higher connectivity in younger adults was associated with better performance on motor and working memory tests. Larger cerebellum volume was also associated with better cognitive performance. When cerebellar activation during cognitive tasks was measured by fMRI, older adults showed decreased activation compared to younger adults. These findings suggest that (1) older adults may be less able to offload cortical processes to the cerebellum due to degraded connections and (2) intact scaffolding may be an indicator of aging resilience.

Along with differences in age, several studies have demonstrated sex-related differences in cerebellar structure in older adulthood. Compared to males, females experience more negative age-related impacts on global cognitive scores and have a higher estimated lifetime risk of AD. Dr. Bernard’s group found that, from approximately age 30 to 40, right posterior cerebellum volumes were larger in females than in males, but experienced a relative decrease among females after age 50. The group also found that hormonal transition periods (e.g., puberty, menopause) are associated with changes in grey matter and that cerebellar functional connectivity patterns vary with hormone levels across the menstrual cycle. For example, females using estrogen supplement therapy in later life showed lower decreases in cerebellum and prefrontal cortex volume, indicating a neuroprotective effect of estrogen in aging.

To investigate the relationship between cerebello-cortical connectivity and hormonal patterns, Dr. Bernard’s group analyzed data from the Cam-CAN repository. The group identified four different hormonal stages in females: reproductive, perimenopausal, early postmenopausal, and late postmenopausal. Analysis revealed progressive reductions in cerebello-cortical connectivity from the reproductive stage through the late postmenopausal stage, particularly in Crus I and II and lobules V and VI. Furthermore, overall lobular cerebellar volume showed progressive reduction, although these reductions were not observed in all individual lobules. When females were compared to age-matched males, males showed fewer age-related differences in cerebello-cortical connectivity.

Dr. Bernard’s group is currently conducting a longitudinal study that collects hormonal assay data on males and females: females provide a menstrual diary, and all participants complete a robust behavioral battery, sleep tracking, questionnaires, and neuroimaging. The group has begun to identify sex differences associated with age and cerebello-cortical connectivity. Frontal lobe connectivity was higher in postmenopausal females than in males, which may reflect “instability” after hormonal shifts and indicate pathological changes. In addition, 17-beta-estradiol, a predominant estrogen in females, is positively associated with motor network connectivity and sleep levels. Over time, the study will track changes between hormone-brain and behavioral relationships.

Theory and Neuroscience of Cerebellar Cognition and Emotion: Clinical Insights and Implications for Neuropsychiatry Across the Age Spectrum

Jeremy Schmahmann, M.D., Harvard Medical School

The cerebellar cognitive affective syndrome (CCAS) is characterized by dysregulation of executive functions (e.g., planning, abstract reasoning, working memory), spatial cognition (e.g., visual spatial organization and memory), language deficits (e.g., aprosodia, anomia), and personality change (e.g., blunting of affect, inappropriate behavior). Dr. Schmahmann’s group defined this syndrome in adults with cerebellar injury, including their proband, a 23-year-old woman with a cerebellar ganglioglioma. They then described CCAS in children following cerebellar tumor resection and noted cognitive challenges as well as affective dysregulation such as irritability, impulsivity and mood swings as a major feature of the presentation. Dr. Schmahmann reviewed the concept of developmental CCAS in twins born very premature (24 weeks) with motor, cognitive, and social-emotional challenges, and cerebellar structural changes on MRI at 4.

The locations of lesions in cerebellar stroke patients determine whether the deficits are motor or cognitive. Whereas lesions in the sensorimotor cerebellum cause dysmetria of movement (the cerebellar motor syndrome), lesions in the cognitive-limbic cerebellum cause dysmetria of thought, i.e., the CCAS. Individuals experiencing dysmetria of thought have impairments in their ability to optimally harmonize behavior according to context, resulting in erroneous social and cognitive responses to the environment, analogous to overshoot and undershoot in the motor system. In one recent case, a 72-year-old man exhibited depression and vestibular symptoms following focal hemorrhage in the deep left cerebellar hemisphere, and FDG-PET showed decreased metabolism in contralateral medial and orbital prefrontal cortex. His symptoms were improved by targeting medial (not dorsolateral) prefrontal cortex with transcranial magnetic stimulation (TMS). These symptoms are conceptualized within Schmahmann’s dysmetria of thought theory which is based on two complementary anatomical realities: (1) the essentially invariant architecture of the cerebellar cortex which enables what he termed the universal cerebellar transform, and (2) the topographic arrangement of cerebellar connections.

The cerebellum is topographically linked in a bi-directional manner through feedforward and feedback loops with most areas of cerebral cortex as well as hypothalamus and ventral tegmental area, and it has multi-synaptic connections with the basal ganglia. The cerebellar corticonuclear microcomplex with its unique computation (the universal cerebellar transform) is a critical node in the precisely arranged distributed neural circuits subserving behavior. In humans, task-based and resting state fMRI studies identified cerebellar regions that regulate motor, language, spatial, and working memory processes, and physiological studies using TMS and resting state fMRI indicate that subregions of the cerebellum control the temporal dynamics of their interconnected cerebral networks.

Aging is associated with decline in conditioned eye-blink responses in both rabbits and humans and decreased numbers of cerebellar Purkinje cells (PCs). In addition, deficits in processing speed, memory, language, and visuospatial and executive functions increase with aging. AD and frontotemporal dementia (FTD) can be differentiated by characteristic network-selective vulnerability in the cerebellum. However, the cerebellum’s contribution to these age-related changes and the relation of these changes to CCAS remains unclear. Researchers should consider using TMS to enhance cerebro-cerebellar interactions, improve cognitive function, and investigate cerebellar contributions to age-related changes.

Resilience and Vulnerability: Neuropathological Measures of Human Cerebellum Morphology and Circuits in Aging and Dementia

Erik Carlson, M.D., Ph.D., University of Washington

Cerebellar brain volume accounts for a large proportion of cognitive performance variance in older adults. Changes in intrinsic cerebro-cerebellar networks are associated with reduced performance during cognitive tasks and are often associated with dementia. However, amyloid plaques and tau tangles are not typically observed in the cerebellum during the early stages of AD.

Dr. Erik Carlson’s group studies the cerebellum’s role in aging and dementia because many anti-amyloid therapies have failed to demonstrate efficacy in clinical trials. Furthermore, regional brain hypometabolism observed in AD cases is not associated with regional amyloid plaque burden. The group collaborated with the large population cohort Adult Changes in Thought (ACT) Study to obtain posterior fossa and cerebellum samples. Using a corrected multivariate regression, the group analyzed dementia status and other dementia-related measures (e.g., Braak and Thal stage, Consortium to Establish a Registry for Alzheimer’s Disease (CERAD) score, age of death, posterior fossa weight) and found that higher posterior fossa weight was associated with protection against dementia. Based on these findings, Dr. Carlson’s group hypothesized that the cerebellum should show neuropathological changes related to dementia such as synapse loss and cell death. Using an initial sample of 30 brains from the ACT Study, the group measured PC linear density in the cerebellum and found no significant differences in density by Braak stage, Thal phase, or dementia status. However, the group evaluated the molecular layer where amyloid plaques often reside in late-stage AD and found that an increased ratio between the widths of the molecular and granular layer is associated with dementia-related differences in Braak stage, Thal phase, and dementia status. These findings suggest that the granular layer neuropil increases during dementia.

To evaluate whether cerebellar synaptic loss is present in dementia, Dr. Carlson’s group stained vesicular glutamate transporter type 1 (Vglut1) in mossy fiber projections from the pontine nuclei to the cerebellum. No significant dementia-related differences were observed in the small puncta density, but the large puncta density decreased with positive dementia status and increasing Braak and Thal stages. By contrast, Vglut1 staining in the molecular layer revealed increases in Vglut1 intensity associated with dementia-related difference, suggesting a potential compensatory change associated with dementia. Dr. Carlson’s group concluded that dementia patients have lower cerebellar volume, reduced granular layer neuropil, and loss of excitatory synapses in mossy fiber rosettes (i.e., the sole input from ponto-cerebellar communication).

The group plans to focus on lateral vermal sections in future studies because it has collected all existing data from dentate nucleus samples. The group has also started conducting spatial transcriptomics and proteomics to analyze dementia-related differences in RNA and protein expression across the cerebellum. Early analysis shows that colony-stimulating factor 1 receptor (CSF1R) is upregulated in dementia samples, indicating microglial activation.

The Cerebellum as a Potential Treatment Target for Late-life Cognitive and Mood Disorders

Vonetta Dotson, Ph.D., Georgia State University

In the United States, 10 percent of adults over age 65 are diagnosed with dementia, and another 22 percent are diagnosed with mild cognitive impairment (MCI). Although 1 to 5 percent of older adults are diagnosed with major depressive disorder (MDD), 10 to 30 percent exhibit subthreshold depressive symptoms that are often associated with subsequent cognitive decline and dementia. Depression and cognitive deficits have most often been studied and targeted for research and treatment within the cognitive control, default mode, and affective networks. However, Dr. Vonetta Dotson’s group has found increased cerebellar blood flow in individuals aged 51 and older with subthreshold depressive symptoms. In addition, higher total volume of the cerebellar vermis 6 was correlated with higher scores on the Center for Epidemiologic Studies Depression (CES-D) total score and somatic symptoms subscale. Other studies have found that individuals with subthreshold depression showed enlarged cerebellum volume, whereas individuals with MDD have smaller cerebellar volumes than healthy controls. These findings suggest that initial neuroinflammatory responses may increase cerebellar volume, but long-term symptoms lead to neuronal degradation and a smaller cerebellum.

Only one-third of individuals with late-life depression respond to traditional depression treatment, and cognitive deficits and brain alterations can persist even when individuals show mood improvement with traditional treatment. These findings indicate that conjunctive or alternative treatments are needed to mitigate depression symptoms and decrease the risk of dementia. Several studies have found that exercise can promote neurogenesis, increase brain volume, decrease AD pathology, improve blood flow and brain metabolism, and improve white matter integrity and functional connectivity in both the cerebral cortex and cerebellum. One study of the cerebellum found that a dance exercise intervention for older adults increased fMRI cerebellum activation during motor tasks and improved performance on logical and verbal memory tasks. Another study found that high intensity interval training led to increased functional connectivity between the cerebellum and both the hippocampus and amygdala.

Recently, researchers have focused on cerebellar neurostimulation therapies for a variety of psychiatric conditions. Several cognitive studies showed mixed outcomes, although a meta-analysis found some benefits from transcranial direct current stimulation (tDCS). Topics for future intervention studies targeting the cerebellum include (1) larger exercise and neurostimulation studies with more diverse participants, (2) exercise and neurostimulation combination therapy, and (3) analysis of treatment response moderators (e.g., demographic characteristics).

Panel Discussion and Q&A

Moderator: Jessica Bernard, Ph.D., Texas A&M University

Participants discussed whether amyloid plaques directly cause neurodegeneration or diffuse toxicity that may affect the cerebellum. Many animal models for ataxia show minimal cell loss in the cortex, indicating that other pathologies underlying ataxia may have cerebellar contributions. Although the cerebellum does not show pathological markers such as amyloid plaques and tau tangles, cerebellar synapse loss is correlated with cognitive deficits and dementia. Researchers may also consider evaluating zebrin (i.e., a protein expressed uniquely in PCs) banding across various cerebellar functional zones to identify more-detailed disease states of cognitive decline and dementia.

Participants agreed that computational models are not necessary for the development of clinical hypotheses and studies, although they can advance understanding of the cerebellum and targeted interventions. Participants also posited that using neurostimulation to target the cerebellum may broadly improve cortical network activities and integration because the cerebellum has diverse outputs to the cortex. Developing better methods for targeting specific areas of the cerebellum can advance the understanding of cerebellar contributions to aging. For example, clinicians can analyze structural MRIs to target specific cerebellar regions for stimulation in individual patients and evaluate PC function to improve the application of electrical fields to those cells. In addition, comparing the efficacy of exercise and targeted neurostimulation in improving cerebellum connectivity may inform how patients can best improve cerebellum connectivity without neurostimulation, which can be expensive and inaccessible to patients.

It is not well understood why children experience more profound impairment from cerebellar insults than adults. Many children with cerebellar injury exhibit delayed language learning and persistent gross language issues, whereas adults exhibit more social and meta-linguistic issues. Moreover, children up to age 10 with normal language skills often exhibit significant loss of language skills following a cerebellar insult. Participants suggested that adults may exhibit less impairment because of compensatory mechanisms within cerebro-cerebellar connectivity that children have not yet developed. However, participants noted the difficulty of developing outcome measures to quantify cognitive deficits and observed that such outcome measures may not sufficiently capture cognitive deficits in adults.

Session 2: Preclinical Evidence of Non-motor Cerebellum Functions

Session Chair: Sam Wang, Ph.D., Princeton University

The Cerebellum as a Context Machine for Action and Thought

Sam Wang, Ph.D., Princeton University

The cerebellum receives short-timescale feedback from other brain regions through an orderly input structure highlighted by the banding of zebrin protein expressed in cerebellar PCs. The mossy fibers of the pons provide inputs to many cells within the granule cell layer of the cerebellum; these cells then converge their outputs on a smaller population of PCs. In addition, climbing fibers of the inferior olive also converge their inputs on PCs to provide information as teaching signals. Not only does the cerebellum learn and process information from climbing fibers, but the cerebellum also relays its own teaching signals; PCs provide information to the thalamus and other cortical regions from cerebellar and other nuclei. Animal studies can advance understanding of these signaling pathways by revealing circuit- and cellular-level cerebellar functions, including (1) the directionality of signaling effects through viral tracing techniques, (2) circuit- and cell-type specificity, and (3) perturbational effects on neural activity, tasks, and behavior that may be beneficial for recovery of cerebellar function.

Researchers can survey the entire cerebellum by using viruses that cross synapses. By injecting these viruses into multiple rodent brain regions, researchers can use whole-brain light-sheet microscopy to identify the directional connectivity between the cerebellum and the neocortex. At the same time, chemogenetics can be used to impair regions of the cerebellum that might inform the association between circuit-level functions and behavioral changes. For example, mice with an impaired lobule VI exhibited difficulties in reversal learning (e.g., the ability to switch paths in a maze after a maze section has been barricaded). Furthermore, these impaired mice did not adapt their habituation behaviors: whereas healthy controls adapted to traversing the perimeter of a novel enclosed arena by day 2, impaired mice continued traversing the middle of the arena. Light-sheet microscopy of cerebellar c-Fos, which is sensitive to neural activity, revealed that disruption of lobule VI inhibits a thalamus-centered network for reversal learning. These findings suggest that lobule VI plays a role in thalamic and brain-wide coordination of learning behaviors. In addition, activation of Crus I PCs via optogenetic techniques accelerated learning of a work-memory task in rodents. This accelerated learning may occur because Crus I enhances the salience of sensory stimuli in the cerebellum, which can help focus rodents on tasks.

Animal studies can support analysis of the cerebellum as a source of priors (i.e., contextual top-down predictions of interaction), and aid research that investigates whether priors are taught within the cerebellum or whether the cerebellum teaches other cortical regions. For example, cerebellar injury at birth has a large risk ratio associated with the development of ASD, which may occur because the cerebellum teaches priors to other cortical regions during sensitive development periods and brain maturation. When the tuberous sclerosis complex 1 (Tsc1) gene—which is highly correlated with the development of ASD—was knocked out of rodent PCs, dendritic spines of pyramidal neurons in the neocortex increased. Researchers were able to image these dendritic spines in vivo using chemical clearing techniques that cannot be applied in humans. In vivo techniques like this can support research on the cerebellum’s role in ASD development and in teaching priors during brain maturation.

Social Memory Deficit Caused by Dysregulation of the Cerebellar Vermis

Yi-Mei (Amy) Yang, Ph.D., University of Minnesota

The cerebellum—particularly the cerebellar vermis—is part of the limbic system, which includes the amygdala, hippocampus, and prefrontal cortex. Through extensive connections with the forebrain, the cerebellum is conceptualized as an estimator of internal states for the operation of movement, emotions, and cognition which collectively affect social behavior. Social behavior includes social cognition (i.e., the ability to mirror others’ actions and to understand the mental states of oneself and others) and social memory (i.e., the ability to distinguish familiar from novel conspecifics by recalling previous encounters). The role of the cerebellar vermis in social behavior is supported by clinical observations of vermal pathology or dysfunction in neuropsychiatric disorders characterized by social behavioral deficits such as ASD.

Dr. Yang’s lab aimed to explore whether and how changes to the cerebellum alone may lead to social behavior deficits. Several ASD mouse models show reduced cerebellar output activity from PCs because of increased excitability in molecular layer interneurons (MLIs) that suppresses PC firing. To control specific cerebellar subregions, the lab selectively increased MLI excitability via a chemogenetic approach in the cerebellar vermis IV/V or vermis VI/VII. In a two-trial social recognition experiment, all mice, with or without the vermal manipulation, explored a social stimulus (stranger mouse) more than a non-social stimulus (empty cup), demonstrating normal sociability. However, when a novel stranger was introduced into the arena, control animals explored the novel stranger more than the familiar one, whereas the vermis-perturbed animals did not show such a preference, exhibiting poor social recognition indices. The researchers repeated the experiment with inanimate objects and found that all groups preferably explored the novel object over the old one. These findings suggest that excitation of MLIs in the vermis does not affect sociability or object recognition memory but does disrupt social recognition memory.

Dr. Yang’s lab implemented an optogenetic approach to investigate the stage at which the cerebellum engages in social memory processing (i.e., encoding, storage, or retrieval of social information). The researchers repeated the social recognition test while delivering MLI photostimulation during the first trial. This manipulation did not affect animals’ social preference or social memory. In contrast, when the photostimulation was delivered during the second trial while a novel stranger was present, unlike the control animals, the vermis groups did not prefer the novel stranger, indicating an inability to recall previous encounters with the familiar mouse. These results suggest that the cerebellar vermis plays a specific role in the retrieval of social memory, but not its encoding.

Following the social recognition test, the Yang lab analyzed c-Fos expression as a cellular marker for neuronal activity in multiple brain regions highly relevant to social behavior, finding that c-Fos levels were sensitive to vermal manipulation. Analysis of distinct patterns of interregional correlation matrices revealed that perturbing the vermis reduced brain-wide connectivity and disorganized neural network structure to form an amygdala-centered network. The lab explored this network further by injecting an anterograde viral vector into the cerebellar fastigial nucleus and a retrograde viral vector into the basolateral amygdala to simultaneously trace the cerebellar outputs and amygdalar inputs. This step uncovered a map of amygdalar integration in the cerebello-cortical circuits. Collectively, these results suggest that the cerebellum co-activates and coordinates the limbic structures essential for emotional responses and cognitive functions to support social memory. The Yang lab plans to conduct future studies using in vivo electrophysiology and calcium imaging to address how the cerebellum interacts with other regions in a task-dependent manner. The group also plans to explore alterations of the cerebello-cortical networks in AD animal models.

Reward Signals in Cerebellar Circuits

Michael Hausser, Ph.D., International Brain Laboratory

The cerebellum is an ideal region for understanding the link between neural circuit activity and behavior because cerebellar circuits have a uniquely simple organization consisting of cellular building blocks. While these circuits are thought to employ forward models of learning in motor function, characterization of their functional organization is limited. Dr. Michael Hausser’s lab investigated how the activation of PCs changes with learning and during adaptation to perturbations in a murine model of sensorimotor integration.

Mice were trained in a novel virtual reality task involving movement of an on-screen object to a target in order to receive a reward, which over multiple days of training led to an improvement of task performance (classified as “slow learning”). When the team applied gain-scaled increases to the object’s trajectory speed, mice who had achieved expertise in the task quickly adapted to the perturbation (demonstrating “fast learning”) compared to naïve or learning mice. Dendritic calcium signals measured from GCaMP6f expression in PCs of expert mice were quantified to observe the signaling mechanisms driving this adaptation. PCs exhibited low-frequency calcium signals that corresponded to activation of climbing fiber inputs, which according to the conventional model deliver motor error signals to the cerebellum. They demonstrated that it is possible to monitor the climbing fiber signals in the same Purkinje cells in lobules V and simple over the timescale of several weeks, allowing them to monitor how these signals change in individual cells during the course of fast and slow learning.

PCs only rarely expressed error signals consistent with forward model activation theories. Instead, the team found robust signals associated with the reward administered during the tasks. Reward signals were more prominent than motor signals in lobule simplex PCs, whereas lobule V PCs exhibited primarily signals associated with motor activation, indicating functional differentiation between cerebellar regions. Further clustering analyses of the activated neurons identified functionally defined PC populations in regionally distinct microzones which could be further categorized as either reward-activated or reward-suppressed. Microzones expressed significantly stronger activation in both categories for unexpected rewards than expected ones, and naïve mice exposed to the task showed signal suppression for expected rewards as they were trained.

By providing evidence of functionally distinct microzones in the cerebellum, Dr. Hausser’s team’s findings complement other recent studies showing reward signaling in different elements of the cerebellar circuitry, and support the longstanding hypothesis that the cerebellum is involved in cognitive processing. The reward signals Dr. Hausser’s team observed were similar to dopaminergic signaling in the ventral tegmental area (VTA) with only minor differences (e.g. that dopamine neurons and Purkinje cells report signed vs. unsigned prediction errors, respectively). Dr. Hausser’s lab is currently using an anatomical tracing approach to identify and compare the input and output pathways of lobules V and simplex to determine whether upstream and downstream signaling occurs in segregated or integrated circuitry. Initial results support an integrated signaling pathway, with some degree of segregation in reward signaling inputs. There was no evidence of direct VTA input to climbing fibers of the cerebellum, suggesting that the reward signals mediated by climbing fiber activation do not result from monosynaptic VTA input. Future work should address the anatomical source of these reward pathways and continue to differentiate the sources and outputs of cerebellar reward circuitry to improve our understanding of how the cerebellum guides cognitive processing.

Cerebellar Modulation of the Brain’s Dopaminergic Circuitry

Kamran Khodakhah, Ph.D., Albert Einstein College of Medicine

Anatomical investigations of the cerebellum reveal several projections to cerebral dopaminergic neurons that regulate cognitive and motivational behavior, including the VTA and substantia nigra pars compacta (SNc). Dr. Kamran Khodakhah’s lab focused on the modulatory effects of these projections on downstream targets of the VTA and SNc using optogenetic activation of projections from deep cerebellar nuclei (DCN) in mice.

Some dopaminergic neurons in the VTA project to the nucleus accumbens (NA), driving reward-associated behaviors like addiction. Dopaminergic VTA pathways also output to the prefrontal cortex (PFC) and are thought to mediate decision-making and other cognitive behaviors. The cerebellar-VTA pathway appears to support these cognitive functions: inhibition of this pathway during social tasks results in behavioral impairments. Optogenetic activation of the cerebellar-VTA pathway rapidly increased synaptic firing in the medial prefrontal cortex (mPFC) as well as NA activation, suggesting that the cerebellum can modulate the VTA and its dopaminergic targets. A second study of cerebellar-mediated dopamine release in VTA targets found a faster time course for release in the NA than in the PFC. This finding aligns with observed cerebellar input on how the VTA typically acts upon these regions.

Dr. Khodakhah’s lab also examined projections from DCN neurons to the SNc, which targets movement-related neurons in the striatum. All three deep cerebellar nuclei send monosynaptic projections to 50 percent of the SNc, roughly dividing half of these projections onto dopaminergic neurons and half onto GABAergic neurons. Optogenetic stimulation of cerebellum-SNc projections increased striatal dopamine levels, suggesting that this pathway may have a functional role in cerebellar modulation of the basal ganglia. DCN-SNc projections might also contain movement-related information: activity in these neurons correlated with SNc dopamine signaling that directly preceded motor behavior in murine models of locomotion. Dr. Khodakhah’s lab also found evidence of bilateral reward-related activity within the same projections. These neurons may modulate vigor to drive motor execution, but the definition of vigor must be better operationalized for future studies. Dr. Khodakhah’s findings support a model of cerebellar function that performs similar roles in both cognitive and motor circuits. Although further research is needed to characterize specific excitatory and inhibitory modulations, the cerebellum appears to drive a mechanism for learning in both motor and non-motor circuits through feedforward processing.

Panel Discussion and Q&A

Moderator: Sam Wang, Ph.D., Princeton University

While several dopaminergic regions in the cerebrum have been linked to projections from the cerebellum, it is unknown whether the cerebellum processes a uniform command transform for all connected regions or whether it executes specialized processes for each specific region. Participants discussed the implications of overlap between cerebello-cortical dopaminergic projections, noting that only 5 percent of observed projections from the DCN projected to both the SNc and VTA. Participants then discussed the definition of vigor in this context. Some Parkinson’s disease (PD) case studies report preserved reaction time in patients with high motivation, such as patients evacuating a room during a fire alarm. While participants acknowledged that such case studies should not be accepted as a universal trend, they proposed that vigor might be defined as motivation to execute behavior. The loss of vigor could provide a neurobiological mechanism for fatigue, which PD patients frequently report. The cerebellar projections to the SNc also suggest cerebellar involvement in the progression of PD itself—fMRI studies of PD patients reveal increased cerebellar resting state EEG, which could potentially indicate compensatory measures for PD-related signal loss in the SNc.

Studies of cerebellar connectivity and cognition are often conducted using animal models. However, reward behavior paradigms using animal models present challenges when translated to human behavior. Participants acknowledged that researchers must carefully distinguish animal models’ reward behavior from the licking behavior expressed during food rewards. Reward behavior in animals often correlates with food-seeking behavior, whereas humans are capable of conceptualizing more abstract rewards (e.g., money). Some fMRI data suggest that the human cerebellum responds to reward cues in the same way that animal models do, but further work must investigate whether this activity is observed throughout reward-associated regions of the brain. c-Fos expression suggests that the cerebellum may coordinate cognitive reward responses throughout these regions in a task-dependent manner, aligning with accepted theories of the cerebellum’s role in coordinating motor behaviors.

Meeting Summary | Day 2

Session 3 | Cerebellar Roles in Affective Processing, Social Function, and Cognition

Session Chair: Krystal Parker, Ph.D., University of Iowa

Cerebellar Circuits, Timing, and Cognition

Krystal Parker, Ph.D., University of Iowa

Abnormalities in connections between cerebellar areas and the frontal cortex have been implicated in cognitive dysfunction. Studies suggest that cerebellar stimulation therapies can rescue these connections. Dr. Krystal Parker’s lab examined the effects of transcranial stimulation on cerebello-frontal circuitry in patients with schizophrenia or bipolar disorder (BD), focusing on a possible cognitive pathway between the dentate nuclei, thalamus, and PFC. Characterizing the network between these regions could inform cerebellum-targeting therapeutics for cognitive dysfunction, especially for behavioral processes that typically decline during aging or neurodegenerative disease progression.

Dr. Parker’s team examined cerebellar cognitive function using human and rodent models of interval timing, a cognitive process known to decline with age. Patients with schizophrenia expressed more variance and earlier predictions of interval times than healthy control groups; patient EEG recordings also showed attenuation of a low-frequency oscillation rhythm observed in control patients before a timing cue. Mice infused with a unilateral D1 dopamine receptor (D1R) blocker in the medial frontal cortex (MFC) also expressed timing behavior similar to subjects with schizophrenia, with the same attenuation of low-frequency rhythms. Low-frequency optogenetic stimulation of the cerebellar terminals in the ventrolateral thalamus rescued timing behavior, suggesting that the cerebellum might regulate low-frequency oscillation rhythms that regulate cognitive functions like timing.

To characterize the utility of these low-frequency rhythms as clinical treatments, Dr. Parker’s team investigated the effects of cerebellar stimulation therapies on patients with mood disorders. Patients with either BD or schizophrenia show evidence of abnormal D1R expression; BD and schizophrenia patients also appear to express similar interval timing behavioral impairments. Patients with PD or other age-related diseases commonly associated with dopamine loss also demonstrate impaired timing behavior, providing further evidence that a D1R stimulation pathway contributes to the regulation of cognitive behavior. Dr. Parker’s lab had previously found no effects on timing in a trial of transcranial pulsed current stimulation (tPCS) on schizophrenia patients despite reinstating low-frequency rhythms. However, this could suggest that tPCS-induced changes were not sufficient to alter brain plasticity and influence cognition. Dr. Parker’s lab is analyzing self-reported and neurophysiological measures of mood disorders in cohorts of patients with schizophrenia and BD following cerebellum-focused TMS therapy. The lab is also investigating the role of the thalamus in driving D1R dopaminergic regulation. It has identified two classes of neurons, tonic and bursting, in the pathway from cerebellar nuclei to the thalamus and frontal cortex. The team plans to characterize these further in future work.

Computational Modeling of Cerebellar Stimulation Using Lobule-Specific Dosage and MRI-Based Approach

Zeynab Rezaee, Ph.D., NIH/NIMH

Medical interventions have used cerebellar stimulation for several centuries to restore behavior, focusing in recent years on non-invasive therapeutics like tDCS. Cerebellar volume is a better predictor of cognitive decline in older adults than PFC volume. Thus, the cerebellum is of particular interest in therapeutic interventions for aging. However, these non-invasive treatments often use a generalized protocol for patients that fails to account for interindividual variability. Age-related differences in brain volume and general anatomical factors vary significantly between adults, contributing to variance in the focality and directionality of administered electromagnetic frequencies (EF). Dr. Zeynab Rezaee’s lab examined variability in treatment efficacy by comparing tDCS buccinator muscle montages across the human lifespan, including in cases of neurodegeneration. Age-related increases in CSF volume dispersed more of the tDCS current, showing altered EF directionality in older adults. The team also observed reductions in frequency magnitude concurrent with age-associated reductions in cerebellar volume. This finding indicates that therapeutic stimulation might have reduced efficacy for older adults already experiencing cerebellar deficits.

To address interindividual variability during tDCS treatment, Dr. Rezaee and colleagues proposed the Cerebellar Lobules Optimal Stimulation (CLOS) system which uses a combination of computational models and experimental validation to improve treatment specificity and reduce variability. Dr. Rezaee’s lab ran simulations on participants’ isolated cerebellar MRI data using the CLOS system to predict the effects of tDCS treatment. These simulations demonstrated significantly higher frequency magnitudes for the optimized model. These results can be validated using additional imaging techniques and behavioral studies. Dr. Rezaee additionally identified some barriers to using the CLOS system. For example, clinicians may not have sufficient resources for individualized treatment. However, the CLOS system demonstrates an effort to optimize stimulation therapies and highlights the need for further research on the unique morphometry of the cerebellum.

When asked if this optimization could be applied in a real-time closed loop experimental protocol, Dr. Rezaee agreed that recent advances in machine learning make such a protocol possible, although these functional assays are not within the scope of her lab’s current work.

Specifying Constraints on Cerebellar Contributions to Prediction

Richard Ivry, Ph.D., University of California Berkeley

A core question concerning cerebellar function is whether the cerebellum performs a common computation that is invoked across multiple tasks or whether cerebellar computations are heterogenous and tailored to various tasks. The former approach is known as the uniform cerebellar transform hypothesis. Various UCT hypotheses have been proposed, including the idea that the cerebellum operates as a forward model to predict future states. In the motor domain, the forward model idea has provided an account of the role of the cerebellum in sensorimotor adaptation. Human subjects demonstrate age-related differences in sensorimotor task performance: older adults show stronger implicit adaptation show a marked reduction in explicit, strategy-based learning. Turning to cognition, Dr. Richard Ivry’s lab has examined the strengths and limitations of prediction-based modeling. In a semantic prediction task, ataxic and control patients were equally sensitive to prediction-gain benefits from sentences with higher predictability. Thus, Dr. Ivry argued that prediction might be too broad a concept for a uniform cerebellar transform; it is important to specify constraints on cerebellar prediction in motor and non-motor domains.

Dr. Ivry’s lab proposed that mental transformations lie on a continuum of dynamic to static predictions called a Continuous Representational Transformation (CoRT). Dynamic transformations are believed to be cerebellar-dependent and driven by gradual and continuous changes in mental representations that can be simulated to predict future states. In contrast, static transformations are cerebellar-independent and may be discretely modified based on learned associations between successive static states. Evidence for the CoRT model comes from observed coordination deficits rather than task selection deficits in patients with cerebellar degeneration, as well as dynamic behavior observed in both motor-focused sensorimotor tasks and non-motor duration perception tasks. To directly test the CoRT hypothesis, Dr. Ivry’s lab conducted a series of visual cognition, arithmetic, and language tasks to observe how ataxic and healthy participants responded to dynamic and state transformations. Ataxic groups showed CoRT-dependent performance declines in both the dynamic visual cognition and arithmetic tasks, whereas their performance in the static tasks was comparable to that of healthy controls. However, ataxic patients and controls performed similarly on both dynamic and static language tasks with no interaction effects. These findings could suggest that language relies on preserved memory retrieval rather than actual prediction, which will be studied in future work.

The theory of uniform cerebellar transformation is supported by the homogenous architecture of the cerebellum, which provides a feasible organization for such a model, as well as evidence from motor, visual cognition, and arithmetic tasks that suggest transformation across motor and non-motor domains. Future work should address CoRT-differentiated tasks of cognitive control, spatial attention, social cognition. Dr. Ivry’s lab aims to further specify constraints on continuous and temporal transformations.

The Human Cerebellum in Social and Emotional Processing

Ingrid Olson, Ph.D., Temple University

In an exploratory application of an algorithm for resting state MRI searchlight analysis, Dr. Ingrid Olson’s team discovered major differences in cerebellar representations for a dataset of subjects with ASD. Despite being the second-highest risk factor for ASD diagnoses, the impacts of cerebellar insults on social aspects of ASD are not well-characterized. Using a combination of meta-analyses and tractography, Dr. Olson and colleagues investigated the effects of age and cerebellar insult location on ASD-typical social behavior.

Although children who sustain cerebellar injury show high rates of altered social behavior, the impact of cerebellar insult in older adults is less consistent. Dr. Olson acknowledged that these mixed results may stem from either (1) issues with the ecological validity of tasks used to examine social behavior in adults or (2) biases in how clinical researchers report changes in cerebellum-mediated social behavior. Several theories exist to explain why cerebellar damage-mediated social deficits might vary across the lifespan. One theory suggests that neurobiological differences (e.g., cerebellar reserve) may drive cognitive resistance in older adults, while others propose adjusting outcome measures for social function in older adults to better capture nuances in behavioral alteration. An additional theory suggests that children may rely on a forward model of learning in the cerebellum, which may shift to a more damage-resistant inverse model over the lifespan. In this case, social deficits would likely manifest as slowness or difficulty initiating automated social responses. Dr. Olson emphasized that data on cerebellar damage-mediated social deficits in older adults are limited, exposing a knowledge gap that future work should address.

Dr. Olson’s lab also sought to characterize the regions of the cerebellum involved in social and affective processing. Their analyses found distinct regions of activation between the cognition-associated regions of interest and the more anterior motor areas. Probabilistic tractography found white matter connectivity between mentalizing regions of the cerebrum and the cerebellum, with preference for connectivity between cerebellar Crus I/II and the midline regions of the frontal lobe. Affect and emotional processing were localized in the vermal regions of the cerebellum. Dysregulation of reward networks in the brain are thought to underlie affective disorders because roughly 30 percent of older adults with cerebellar damage experience novel onsets of MDD or BD. Dr. Olson’s team sought to examine how cerebellar damage might impact pathways connecting to reward-associated regions like the VTA. The team used probabilistic tractography imaging to investigate projections from the cerebellar cortex and deep nuclei to the VTA. While all deep nuclei projected to the VTA, interposed and ipsilateral connections were denser. Medial regions of the vermal and paravermal cerebellar cortex also showed higher connectivity with the VTA, aligning with previous studies that found activation of multiple regions of the medial cerebellar cortex during emotion-evoking tasks. Dr. Olson’s lab also found that higher connectivity between the cerebellum and VTA predicted higher anxiety and depression scores in healthy adults, suggesting that there may be an optimal amount of connectivity for mood behavior regulation. Evidence suggests that prodromal cases of AD affect the VTA, although more work is needed to characterize those effects. Dr. Olson’s lab plans to investigate the role played by interactions of the cerebellar-VTA pathway in the larger mesolimbic system and how this pathway may respond to non-invasive stimulation therapies.

Panel Discussion and Q&A

Moderator: Krystal Parker, Ph.D., University of Iowa

Participants noted that patients with BD are more likely to develop posterior fossa arachnoid cysts than healthy controls and that patients with such cysts present with milder symptoms of BD, forming a potential BD subtype. These posterior fossa cysts suggest that the vermis influences the limbic system and may be linked to more vestigial circuits. However, participants also acknowledged the limitations in assigning value to the functional representations of the vermis and cerebellum. Parasagittal subzones are difficult to discern with current imaging techniques and technological advances may be needed to capture the functional role of microzones. Zebrin staining techniques may also be misleading because functional connectivity and functional organization may be exclusive of each other to some extent.

One question addressed the translational applicability of these cerebellar studies and inquired whether the minute strength changes involved in optimizing tDCS would yield differentiable impacts on function. Because current tDCS treatments observe synaptic-level changes in PCs but not deeper-level granule cells, fine-tuning the strength of stimulation may help the signal reach deeper neurons. However, the geometry of the cerebellum makes it difficult to treat a region of interest using only one angle or strength setting. Further clinical work should aim to develop strategized treatment plans for specific regions.

Participants discussed how to better capture the cerebellum’s functional roles. For example, the construct validity of paradigms addressing cerebellar language regions may need improvement. Participants also addressed the automation functions of the cerebellum: they agreed that, although the cerebellum is integral in smoothly transforming a given series of commands, it might not play a significant role in how those commands are chosen. In the aging brain, these commands may be contained within a cortical framework that also supports memory consolidation.

Overall Discussion

Moderator: Coryse St. Hillaire-Clarke, Ph.D., NIA

Participants discussed limitations of fMRI, BOLD, and other imaging techniques in measuring cerebellar function and activity. Functional imaging primarily reveals cerebellar cortical inputs, making it difficult to infer whether the cerebellum is receiving efferent copies or processing and encoding information itself. For example, a BOLD signal response in the cerebellar cortex could occur either from neural activity in climbing fibers or from input from the pons. It is not possible to understand how cerebellar BOLD signals translate to specific instances of neural activity without biophysical maps of neural activity. Participants agreed that conducting secondary analyses using cerebellum images from large repositories could help researchers better correlate neurophysiological studies with functional measures. Repositories such as the Alzheimer's Disease Neuroimaging Initiative (ADNI) and Cam-CAN provide imaging data and a range of other functional measures from older adults.

Participants suggested a three-step process to address how the cerebellum engages with neuromodulatory systems in the brain and relates to brain state dynamics disrupted by dementia: (1) map connectivity of the cerebellum in humans; (2) study how connectivity changes in the context of aging and neurodegenerative disease; and (3) once specific cerebellar changes have been identified, use animal models to genetically manipulate specific circuits to test targeted interventions. Because current rodent models of AD do not reflect important aspects of the disease in humans, researchers should consider beginning cerebellum physiology studies in aged wildtype rodents before identifying or developing transgenic models.

Closing Remarks

Matt Sutterer, Ph.D., NIA

Dr. Sutterer thanked all meeting participants for providing expertise to advance cerebellum and aging research. He encouraged meeting participants to review NIA funding opportunities, as well as the NIH Brain Research Through Advancing Innovative Neurotechnologies (BRAIN) Initiative managed by the National Institute of Neurological Disorders and Stroke (NINDS).

Contact Information

Please contact Matt Sutterer at matt.sutterer@nih.gov if you have questions about the workshop.