Authors: Jessica A. Bernard (1Department of Psychological and Brain Sciences, Texas A&M University College Station, TX 77843-4235, United States; 2Texas A&M Institute for Neuroscience, Texas A&M University College Station, TX 77843-4235, United States)
Categories: Article
Source: Cerebellum (London, England)
Authors: Jessica A. Bernard
The cerebellum is recognized as being important for optimal behavioral performance across task domains, including motor function, cognition, and affect. Decades of work have highlighted cerebello-thalamo-cortical circuits, from both structural and functional perspectives. However, these circuits of interest have been primarily (though not exclusively) focused on targets in the cerebral cortex. In addition to these cortical connections, the circuit linking the cerebellum and hippocampus is of particular interest. Recently, there has been an increased interest in this circuit, thanks in large part to novel findings in the animal literature demonstrating that neuronal firing in the cerebellum impacts that in the hippocampus. Work in the human brain has provided evidence for interactions between the cerebellum and hippocampus, though primarily this has been in the context of spatial navigation. Given the role of both regions in cognition and aging, and emerging evidence indicating that the cerebellum is impacted in age-related neurodegenerative disease such as Alzheimer’s, I propose that further attention to this circuit is warranted. Here, I provide an overview of cerebello-hippocampal interactions in animal models and from human imaging and outline the possible utility of further investigations to improve our understanding of aging and age-related cognitive decline.
Cerebellar contributions to non-motor behaviors are now a well-known and accepted finding in the literature, thanks to several decades of research demonstrating altered cognitive and affective function after infarct (J. D. Schmahmann et al., 2019; J. Schmahmann & Sherman, 1998; Timmann et al., 2008, 2009) and functional neuroimaging showing cerebellar activation present across cognitive, social, and affective task domains (Balsters et al., 2013; King et al., 2019; Stoodley et al., 2012; Stoodley & Schmahmann, 2009a; Van Overwalle et al., 2014, 2020). These non-motor functions are further supported by our understanding of circuitry linking the cerebellum to the cortex (via the thalamus), seen in both human and non-human primates (Bernard et al., 2012; Buckner et al., 2011; Dum & Strick, 2003; Kelly & Strick, 2003; O’Reilly et al., 2010; Salmi et al., 2010; J. D. Schmahmann & Pandya, 1997; Steele et al., 2017; Strick et al., 2009). While much of the focus on the cerebellum in the context of cognition has been related to the networks linking the cerebellum to the prefrontal cortex and has focused on fluid domains such as working memory, there is a growing literature suggesting that other non-motor domains, including spatial navigation and long-term memory may also rely upon cerebellar processing for task performance (Babayan et al., 2017; Bernard et al., 2020; Iglói et al., 2015; King et al., 2019; Rochefort et al., 2013). Such an expansion of our understanding of cerebellar contributions to behavior also in turn suggests that additional circuits may be of particular interest. Here, I suggest that the circuit and interactions between the cerebellum and hippocampus is of particular interest. Work in navigation (discussed further below) has provided a key foundation for this assertion, and has been reviewed recently (Rondi-Reig et al., 2002). I provide an overview of the cerebellum in cognition more broadly, and then discuss cerebello-hippocampal interactions across the animal and human literature. Finally, I suggest that this is a circuit with great potential for advancing our understanding of cerebellar contributions to cognition, as well as to aging and age-related neurodegenerative disease.
The cerebellum is now widely recognized for its contributions across functional domains (Buckner, 2013; King et al., 2019; J. D. Schmahmann, 2018; J. D. Schmahmann et al., 2019; J. Schmahmann & Sherman, 1998; Stoodley & Schmahmann, 2009a), including various components of cognition. Historical conceptualizations of cerebellar function in the last century primarily focused on motor behaviors, in no small part due to the overt deficits seen in individuals with cerebellar lesions (Holmes, 1939). However, as early as the 1980s it was theorized that the cerebellum may contribute to higher cognitive processes (Leiner et al., 1986, 1989, 1991). Shortly thereafter, Schmahmann and Shermann (J. Schmahmann & Sherman, 1998) described the cerebellar cognitive and affective syndrome, wherein individuals with lesions to the posterior aspects of the cerebellum were shown to have deficits in both cognitive and affective processing. This provided striking support for non-motor functions of the cerebellum. Shortly thereafter, foundational work using viral tract tracing methodologies in non-human primates was conducted demonstrating parallel closed-loop cerebello-thalamo-cortical circuits, linking distinct aspects of the cerebellum to prefrontal, motor, and parietal cortices (Clower et al., 2001; Dum & Strick, 2003; Kelly & Strick, 2003). This provided converging evidence in support of non-motor contributions of the human cerebellum. With the advent of human neuroimaging subsequent work investigating cerebellar functional activation across task domains has provided additional support for cerebellar involvement in higher cognitive processing and affect (Balsters et al., 2013; King et al., 2019; Stoodley et al., 2012; Stoodley & Schmahmann, 2009a, 2009b, 2010). Finally, connectivity measures have further provided support for this idea. Diffusion imaging has shown parallel closed-loop circuits in the human brain (Bernard et al., 2016; Rousseau et al., 2022; Salmi et al., 2010; Steele et al., 2017), while resting state functional connectivity has demonstrated cerebellar involvement in broad cortical networks (Bernard et al., 2012, 2014; Buckner et al., 2011; Krienen & Buckner, 2009; O’Reilly et al., 2010). With our understanding of cerebellar function now widely known to include contributions to higher cognitive processing, there is also growing interest in understanding whether and how the cerebellum may contribute to disease and behavioral differences across the lifespan.
Again, thanks to several decades of research there is converging evidence implicating the cerebellum in psychosis (Andreasen et al., 1996, 1998; Andreasen & Pierson, 2008; Bernard et al., 2017; Bernard & Mittal, 2015; Kent et al., 2020; Lundin et al., 2021; Parker et al., 2014), autism (D’Mello et al., 2015; Mostofsky et al., 2009; Stoodley et al., 2017), and attention deficit hyperactivity disorder (Hove et al., 2015; Kucyi et al., 2015), as well as in neurological disorders (outside of those with a cerebellar etiology) such as Parkinson’s Disease (Festini et al., 2015; Gilat et al., 2017; Wu & Hallett, 2013), and in aging (Arleo et al., 2023; Bernard, 2022; Bernard et al., 2013, 2020, 2021; Bernard & Seidler, 2013, 2014). In aging, it has been shown that cerebellar volume is smaller (Bernard & Seidler, 2013; Han et al., 2020; Jernigan et al., 2001; Raz et al., 2010), and there are differences in both functional connectivity (Ballard et al., 2022; Bernard et al., 2013, 2021), and functional activation patterns (Bernard et al., 2020; Jackson et al., 2020). Recent conceptualizations have suggested that the cerebellum serves as a critical source of scaffolding for cortical processing wherein the ability to rely upon cerebellar internal models and more automatic processing is inhibited in advanced age, resulting in a need for increased cortical resources (Bernard, 2022). In advanced age, efference copies are not as effectively being shared with the cerebellum, perhaps in part due to differences in white matter and functional interactions between the cerebellum and cortex. Processing and computations related to these efference copies may also be impacted by the volumetric loss in the cerebellum that is experienced in advanced age. As such the processing of forward and inverse models is negatively impacted in advanced age (Bernard & Seidler, 2014). Subsequently, I proposed that this in turn impacts cortical activation patterns in advanced age, as older adults are less able to rely upon forward and inverse model processing in the cerebellum, as such instead need to recruit additional cortical resources, contributing in part to the bilateral patterns of activation seen in this age group (Bernard, 2022). This work has implications for our understanding of the cerebellum in Alzheimer’s Disease (AD) as well, an emerging area of research (Gellersen et al., 2021; H. I. L. Jacobs et al., 2018; J. D. Schmahmann, 2016).
The role of the cerebellum in AD at this point remains somewhat mysterious, though in the last decade an emerging literature has suggested that the cerebellum may play a role in AD beyond just that of a “silent bystander” (Schmahmann, 2016). That is, in past conceptualizations any cerebellar pathology has been seen as incidental and less critical to our understanding of AD and AD symptomatology (H. I. L. Jacobs et al., 2018). However, it may be the case that while cerebellar pathology is not an underlying cause of the disease, this still impacts function in individuals with AD (H. I. L. Jacobs et al., 2018). In their review, Jacobs and colleagues (2018) provided a needed overview of the current status of the field, and description of cerebellar pathology. This pathology is largely present in early-onset AD, though may also impact individuals with sporadic AD as well. This includes macrostructural changes wherein atrophy is seen in the posterior aspects of the cerebellum and vermis, while amyloid-β is also present in the cerebellum (H. I. L. Jacobs et al., 2018). Further, in the limited literature on functional activation, they note that there appear to be differences in patterns of cerebellar activation when comparing AD patients to both healthy controls and those with mild cognitive impairment (H. I. L. Jacobs et al., 2018). Broadly, they argue that there may be subtle changes in cerebellar computations that impact behavior (H. I. L. Jacobs et al., 2018), like what we previously suggested in typical aging (Bernard & Seidler, 2014). Critically however, this was not a quantitative meta-analysis and many open questions with respect to cerebellar functional activation in AD remain.
In parallel to this review, there has also been a growing literature of careful structural analyses on the cerebellum in AD (Gellersen et al., 2021; Lin et al., 2020; Tabatabaei-Jafari et al., 2017; Toniolo et al., 2018, 2020) as well as investigations demonstrating differences in resting state functional connectivity of the cerebellum (Guo et al., 2016; Olivito et al., 2020). Together, there is an emerging literature that implicates the cerebellum in AD. Given recent conceptualizations that suggest that the cerebellum may be key for supporting cortical function in aging (Bernard, 2022), coupled with the extensive known pathology in the hippocampus in AD, an improved understanding of the cerebellum and cerebellar networks with the hippocampus may provide a more direct circuit by which the cerebellum impacts behavior and symptomatology in AD.
Evidence of interactions between the hippocampus and the cerebellum date back to the 1950s. Work using direct electrical recordings and stimulation in cats demonstrated that cerebellar stimulation results in changes in the electrical activity recorded from the hippocampus, and the cerebellar stimulation reversed seizure activity induced by hippocampal stimulation (Iwata & Snider, 1959). Viral tract tracing methodologies in non-human primate models provided detailed information about cerebellar connections to prefrontal, parietal, and motor cortical regions as well as to the basal ganglia (Bostan et al., 2010; Clower et al., 2001; Dum & Strick, 2003; Hoshi et al., 2005; Kelly & Strick, 2003). Subsequent human work using diffusion tensor imaging has provided converging evidence in the human brain for these parallel closed loop circuits (Bernard et al., 2016, p. 201; Rousseau et al., 2022; Salmi et al., 2010; Steele et al., 2017), and resting state connectivity measures provide broader support for large scale cerebello-cortical interactions that extend across the cortex (Bernard et al., 2012; Buckner et al., 2011; Habas, 2018; Krienen & Buckner, 2009; O’Reilly et al., 2010). This includes contributions across known functional networks and parcellations (Bernard et al., 2012; Buckner et al., 2011; O’Reilly et al., 2010), suggesting that there are likely functional counterparts to much of the cortex and subcortex in the cerebellum. Notably, this also includes functional interactions at rest between the cerebellum and hippocampus (Bernard et al., 2012).
While the density of white matter in the cerebellar peduncles and midbrain can be a challenge for one-to-one mappings of connectivity between the cerebellum and cortex in the human brain, and as such we cannot directly replicate the extensive patterns of cortical connectivity delineated in the resting state, advances in white matter imaging and analysis have certainly allowed for careful investigation of cerebello-thalamo-cortical tracts more generally. As alluded to above, many of the cerebello-thalamo-cortical connections initially traced in non-human primates between the cerebellum and prefrontal and motor cortices have been successfully tracked in the human brain, and researchers have succeeded in parsing tissue in the brainstem (Bernard et al., 2016; Habas & Cabanis, 2007; Salmi et al., 2010). Further, more recently, regional delineations of the dentate nucleus reported in primates (Dum & Strick, 2003) have also been demonstrated in the human brain using diffusion imaging (Steele et al., 2017). However, whether there are structural connections between the cerebellum and hippocampus is an important question. Data from both animal and human samples is of great interest and import.
To this point, there have been indications that the structural anatomy is in fact in place to allow for interactions and connections between the cerebellum and hippocampus. Several studies have successfully tracked white matter from the cerebellum, via the cerebellar peduncles to the thalamus (Bernard et al., 2016; Habas & Cabanis, 2007; Salamon et al., 2007; Salmi et al., 2010), including tracing at the level of the human dentate nucleus (Steele et al., 2017). Furthermore, connections between the thalamus and the hippocampus have also been tracked in the human brain (Behrens et al., 2003; Bernard et al., 2015). However, these connections as described here are from separate studies and instead demonstrate the capacity for hippocampal-cerebellar connections via a disynaptic (at minimum) circuit with the thalamus. There is also limited evidence of connections between these two structures in a study focusing exclusively on this circuit (Arrigo et al., 2013). In an attempt to trace the cerebello-limbic circuit and delineate its existence in the human brain, Arrigo and colleagues completed a study using constrained spherical deconvolution analysis in a small sample of 15 healthy young adults (Arrigo et al., 2014). The authors segmented the amygdala and hippocampus and then completed tractography analyses. The authors suggest direct connections between the cerebellum and hippocampus via the superior cerebellar peduncles (Arrigo et al., 2014), it is not clear how anatomically plausible this is given findings in animal models (Bohne et al., 2019; Watson et al., 2019). It is also important to contextualize the methodology here, as the authors did not use waypoints in their analyses. Such a direct monosynaptic connection is somewhat surprising given the other cerebello-thalamo-cortical circuits that have been mapped in the human (Bernard et al., 2016; Habas & Cabanis, 2007; Salmi et al., 2010; Steele et al., 2017) and non-human primate (Clower et al., 2001; Dum & Strick, 2003; Kelly & Strick, 2003) which suggest disynaptic (at minimum) connections with much of the cortex. The rabies tracing methods in animal models are particularly compelling as they allow for clear delineation of multi-synaptic circuits, while the human tractography has been built upon this original work and primarily relies upon waypoints in the analysis. Thus, while this work provides proof-of-principle evidence of structural connections between the cerebellum and hippocampus in the human brain, it is critical to note that replication of this finding in larger samples using rigorous tractography methods rooted in knowledge of the underlying anatomy based on animal work is critical to validate these results. The suggested monosynaptic direct connection should be interpreted with immense caution, as this is counter to what is seen in animal work.
Animal models have also provided insights into the underlying connections that subserve cerebello-hippocampal interactions. Recent work from Watson and colleagues (Watson et al., 2019) using rabies tract tracing has provided important and needed insights into the structural connections between the cerebellum and hippocampus in the rodent brain. Inspired by the findings from Arrigo and colleagues in the human brain (Arrigo et al., 2014), Watson and colleagues completed tract tracing in rodent models to map the connections between the two structures. Their findings suggested that there are three main inputs to the hippocampus from the inputs from the vestibulo-cerebellum, inputs from central regions of the cerebellum from Vermis VI, and finally, inputs from Crus I via the dentate nucleus. Figure 1 provides a visual summary of their findings with respect to the viral tract tracing data. Notably, they suggest that based on the timing of results from tract tracing these connections are likely multi-synaptic (Watson et al., 2019). Thus, there is converging evidence across species in support of the structural connections that link the cerebellum to the hippocampus. The detailed work here in particular stands to be an excellent starting point for future diffusion imaging work in the human brain, as probable tracts and waypoints are clearly identified in this data. Further, the multi-synaptic nature of these findings suggest that the “direct” results reported by Arrigo and colleagues (2014) in the human brain require careful refinement. While certainly there are white matter pathways that link the cerebellum and hippocampus in the human brain, they are likely to parallel the pathways demonstrated here in rodents (Watson et al., 2019).
In parallel to the rabies tracing, Watson and colleagues (2019) also looked at functional interactions between the two regions. Local field potentials were simultaneously recorded from the cerebellum and hippocampus while the animals where in their home cage environment, and during the learning of goal-directed behavior (crossing a linear track to reach a reward). They demonstrated signal coherence that was in the 6-12 Hz range but also dependent upon behavior (Watson et al., 2019). Together, this provided critical support demonstrating both additional data for a structural basis of interactions between the cerebellum and hippocampus, as well as functional interactions that are behaviorally dependent.
This more recent work from Watson and colleagues (2019) in addition to building on human work, also comes from a foundation of research investigating spatial navigation in rodent models. In a review by Rochefort and colleagues, they outlined the role of the cerebellum in spatial navigation. One key point they made was the suggestion that the cerebellum may in fact provide aid to the hippocampus in the building of spatial maps (Rochefort et al., 2013). This was evidenced by their experimental work with mutant mice that did not have normal cerebellar function. When learning to navigate a water maze, the mutant mice did so poorly as compared to wild type controls. Further, there was disrupted function and firing in hippocampal place cells (Rochefort et al., 2011) in the mutant group. Thus, not only does this work suggest a role for the cerebellum in spatial processing and navigation, but it also suggests that it does so through modulation of the hippocampus, further supporting the links between these two regions. Cerebello-hippocampal interactions have been noted to be especially important for the later phases of learning a spatial sequence when behavior follows the learned sequence of movements (Babayan et al., 2017). A more recent review further underscored these relationships in the context of spatial learning and processing (Yu & Krook-Magnuson, 2015) while highlighting the functional interactions between these regions.
Research on temporal lobe epilepsy has also provided insights into these cerebello-hippocampal interactions, and harkens back to historical work on interactions between these brain regions (Iwata & Snider, 1959). Krook-Magnuson and colleagues investigated whether and how the cerebellum may modulate the temporal lobe and influence seizure activity that originates in the temporal lobe (Krook-Magnuson et al., 2014). Using electroencephalography in the hippocampus in conjunction with an optogenetic based seizure intervention, the authors found further evidence of cerebello-hippocampal interactions. When the optogenetic stimulation targeted lateral regions of the cerebellum, seizures were shortened. This was even more robust when the cerebellar vermis served as the target of the manipulation (Krook-Magnuson et al., 2014). More recently, their research group continued investigations on cerebellar-hippocampal interactions using optogenetics. Optogenetic stimulation was applied to the cerebellum while local field potentials were recorded from the hippocampus and behavior on a spatial memory task was also assessed (Zeidler et al., 2020). Notably, stimulation to the cerebellum impacted neuronal firing in the hippocampus and impacted performance on the spatial memory task, providing important additional support for the functional circuit between the cerebellum and hippocampus, and highlighting its behavioral relevance. Indeed, to highlight the functional importance of this circuit the authors coined the term “hippobellum” to refer to the interactions between the cerebellum and hippocampus (Zeidler et al., 2020). Thus, the limited structural data available to date, coupled with data from rodent models provides compelling evidence for interactions between the cerebellum and hippocampus, such that cerebellar firing may in fact modulate that in the hippocampus. While such causal recording approaches are not readily available in human research (intracranial recording is rare, and often limited in the regions that are able to be investigated), functional imaging has also provided further insights into this circuit.
The work noted above suggests some degree of directionality in that cerebellar stimulation or modulation influences hippocampal function and cellular firing. However, there is also work suggesting influences in the opposite direction. Hoffman and Berry (Hoffmann & Berry, 2009) took the opposite approach in that they looked at hippocampal theta and learning during trace eyeblink classical conditioning in rabbits. Trace eyeblink classical conditioning paradigms involve the presentation of a conditioned stimulus, typically an auditory tone, and there is a delay after the tone terminates prior to the presentation of the unconditioned stimulus, a puff of air to the eye. In human neuroimaging and in animal work it has been demonstrated the trace conditioning involves both the cerebellum and the hippocampus (Cheng et al., 2008; Christian & Thompson, 2003; Woodruff-Pak, Papka, & Ivry, 1996; Woodruff-Pak, Papka, Romano, et al., 1996; Woodruff-Pak & Thompson, 1988). Hoffman and Berry were able to compare conditioning trials that were administered with and without the presence of naturally occurring theta. In the animals with naturally occurring theta, learning was significantly higher, as measured by both the percent of conditioned responses and cumulative conditioned responses. Furthermore there were rhythmic theta oscillations in the cerebellum (both cortex and interposed nuclei) that were synced with the hippocampal theta (Hoffmann & Berry, 2009). In the group of animals where learning was timed without naturally occurring theta, these cerebellar oscillations were not seen. The authors argue that the hippocampus may in fact modulate the cerebellum during this form of learning (Hoffmann & Berry, 2009) based on the group differences in cerebellar theta with and without hippocampal theta. Thus, there is evidence to indicate that there are bidirectional functional interactions between the cerebellum and hippocampus in animal models.
Finally, recent investigations of sleep have also provided insights cerebello-hippocampal interactions (Torres-Herraez et al., 2022). In mice that were sleeping naturally, recordings were collected from the hippocampus and three regions of the cerebellum during different sleep stages (wake, rapid eye-movement; REM, and non-REM) given that both regions are important in learning and memory processes. The recordings indicated that there was coordination in the oscillations in the two regions, though the way the regions were coordinated varied with sleep stage (REM and non-REM). The authors suggest that this may be particularly informative for sleep-dependent memory consolidation (Torres-Herraez et al., 2022), and this also provides further evidence for functional interactions between these two regions. As demonstrated above in eye-blink conditioning (Hoffman & Berry, 2009), these two regions together seem important for learning processes, and while speculative, it may be that these interactions and oscillatory synchronizations may be important for this process.
To this point, functional activations and interactions between the cerebellum and hippocampus have been investigated together using task-based imaging in several functional domains. Evidence for interactions also comes from resting-state functional connectivity. While this work will be discussed further below, across all sources of data, there is a general pattern of findings indicative of key functional relationships between the two regions. Though the methodology is vastly different from that which is applied in animal models, this again provides further converging evidence for cerebello-hippocampal circuits and its role in behavior.
Using a spatial navigation task, Iglói and colleagues (Iglói et al., 2015) demonstrated a functional link between the cerebellum and hippocampus. Navigation was investigated in a sample of young adult males using a virtual star maze that included training, control, and probe trials. They investigated place- and sequence-based responses relative to the control conditions, and found two dissociable circuits associated with Crus I of the cerebellum (Iglói et al., 2015). The circuits notably included frontal and parietal regions, but place-based responses were part of a circuit with the right hippocampus, while sequence-based responses were associated with the left hippocampus. Together, this highlights potential cerebello-hippocampal interactions in the human brain in sequence processing, much like what has been seen in animal models (reviewed in (Rondi-Reig et al., 2022).
Interactions between the cerebellum and hippocampus have also been demonstrated when performing spatio-temporal prediction tasks (Onuki et al., 2015). In young adults that were performing a task made up of timing-dependent finger movements, motor coordination was dissociable from timing (both predictive and reactive), and during predictive timing, there was co-activation of the cerebellum and hippocampus (Onuki et al., 2015). The authors followed-up on these findings by using psychophysiological interaction analysis with seeds defined by hippocampal areas that were active during the predictive timing task and demonstrated bilateral task-based connectivity with the cerebellum. This work moves outside of the spatial domain which has been heavily studied particularly in animal models, and the authors highlight these findings in the context of adaptive timing models (Onuki et al., 2015) suggesting that these interactions are important to multiple domains. Further extending the domains of interaction between these regions is work from Paleja and colleagues who investigated hippocampal interactions during pattern separation using a delayed match to place task modelled after rodent investigations (Paleja et al., 2014). During pattern separation they identified two networks that were associated with the hippocampus. The primary network was made up of additional medial temporal lobe regions, but there was a secondary network including the cerebellum as well as cortical regions across lobes of the brain (Paleja et al., 2014). This provides further support for potential functional interactions between the hippocampus and cerebellum in the human brain.
Finally, resting-state connectivity has also provided some suggestion of functional interactions between the cerebellum and the hippocampus in the human brain. Initial work using independent component analysis to determine unique cerebellar contributions to known cortical networks, found that cerebellar lobule IX in particular contributes to the default mode network (Habas et al., 2009). Notably, this includes the hippocampus and parahippocampal gyrus, suggesting that at rest there are some similarities in the blood oxygen level dependent (BOLD) signal fluctuations, which may be indicative of shared processing. Similarly, in a larger investigation seeking to map known cortical networks to the cerebellar cortex, Buckner and colleagues, demonstrated a connectivity atlas, which again includes cerebellar components associated with broader cortical networks that include the hippocampus and medial temporal lobe regions (Buckner et al., 2011). Using seed-to-voxel approaches that quantify correlations in the BOLD signal from proscribed regions, in this case in the cerebellum, with all of the other voxels in the brain, we investigated differences in cerebello-cortical connectivity in young and older adults (Bernard et al., 2013). Broadly speaking, our work suggested that connectivity in healthy older adults (cognitively normal and in the absence of disease) was lower than in young adults. However, one of the consistent patterns that was revealed across lobular cerebellar seeds was lower connectivity with the hippocampus and paraphippocampal gyrus (Bernard et al., 2013). These age differences suggest that not only are there functional interactions between the cerebellum and hippocampus at rest, as indexed by correlations in the BOLD signal, but that these interactions are negatively impacted by aging. Together, work using resting state connectivity provides further converging evidence to support cerebello-hippocampal interactions in the human brain.
Table 1 provides a summary of the interactions between the cerebellum and hippocampus based on the work discussed here, across human and animal studies. While it should be emphasized that this was not a meta-analysis, there is a notable robust pattern such that cerebellar Lobule VI, and to some extent Crus I appears to be particularly robustly implicated in these connections. This is especially notable in the animal recordings, but also present in the task-based functional connectivity from human neuroimaging. As such, Lobule VI (and to some extent Crus I) may be especially interesting targets for probing this circuit further. However, it is also clear that these interactions generally span the cerebellar cortex and include the deep cerebellar nuclei. Because investigations of this circuit and these interactions are still relatively novel, the field would benefit from careful structural and functional mapping for the cerebello-hippocampal network across species.
As outlined above, there is a growing literature implicating the cerebellum in age-related motor and cognitive declines (Arleo et al., 2023; Bernard, 2022; Bernard et al., 2020; Bernard & Seidler, 2014; Han et al., 2020; Jernigan et al., 2001; Raz et al., 2005), and an emerging literature in Alzheimer’s Disease (Gellersen et al., 2021; Guo et al., 2016; H. I. L. Jacobs et al., 2018; J. D. Schmahmann, 2016; Tabatabaei-Jafari et al., 2017). Coupled with the understanding that the cerebellum contributes to higher order cognitive behaviors (Bernard et al., 2020; King et al., 2019; J. Schmahmann & Sherman, 1998; Stoodley et al., 2012) it is thus of particular interest and importance to consider the interactions between the cerebellum and hippocampus in studies of advanced age.
When looking at the gross anatomy of neuropathology in Alzheimer’s Disease and in aging more generally, the hippocampus is heavily impacted (Adler et al., 2018; Driscoll et al., 2003), though not to the exclusion of other regions. Though the investigation of the cerebellum in Alzheimer’s Disease is increasingly recognized (Arleo et al., 2023; H. I. L. Jacobs et al., 2018), this remains an emerging area of research particularly with respect to neuroimaging. Though there is evidence of the presence of amyloid-ß in the cerebellum particularly in cases of familial disease (Heckmann, 2004; H. I. L. Jacobs et al., 2018; Kim et al., 2021; Sepulveda-Falla et al., 2014; Xia et al., 2022), it is only in more recent years that the field has begun to pay increased attention to the cerebellum thanks in large part to the focused review from Jacobs and colleagues (2018), coupled with novel findings related to cerebellar structure in AD (Gellersen et al., 2021; Guo et al., 2016; Tabatabaei-Jafari et al., 2017). While the cerebellum is not thought to be a primary driver of pathology, it has been suggested to contribute to symptomatology and behavior, particularly for domains that are overlapping with the cerebellar cognitive affective syndrome (H. I. L. Jacobs et al., 2018). In the context of the scaffolding framework of the cerebellum in aging (Bernard, 2022), cerebellar processing and interactions with the cortex are important for optimal function behaviorally, and changes in cerebellar processing may contribute to differing patterns of cortical functional activation in older adults. In the context of AD, this scaffolding role for the cerebellum may influence behavior and symptomatology. If there is relative maintenance and preservation of cerebellar function, this may mask behavioral deficits in the broader context of cortical pathology (though as noted above, the cerebellum is not immune to AD pathology). However, with disease progression such scaffolding may be increasingly less effective. Critically however, this framework did not fully consider interactions between the cerebellum and hippocampus.
As described here, there is strong evidence to indicate interactions between the cerebellum and hippocampus (Krook-Magnuson et al., 2014; Rochefort et al., 2011; Rondi-Reig et al., 2022; Torres-Herraez et al., 2022; Yu & Krook-Magnuson, 2015; Zeidler et al., 2020) and findings from work using animal models indicates that cerebellar activity may directly impact hippocampal function (Krook-Magnuson et al., 2014; Zeidler et al., 2020). While much of the work on these important interactions has focused on spatial navigation (Rochefort et al., 2011; Rondi-Reig et al., 2002, 2022), here, I propose further investigation of this circuit across additional cognitive domains, particularly those related to learning and memory. With respect to memory, recent work in the human brain has suggested that the cerebellum may be involved in this domain (Bernard et al., 2020; King et al., 2019), while the hippocampus is well-recognized for its role in memory processes (Burgess et al., 2002; Henke et al., 1999; Kircher et al., 2008; Lee et al., 2013). With respect to learning, both the cerebellum and hippocampus are conceptualized as working together during motor acquisition (Doyon et al., 2018; Schendan et al., 2003), while classical eye-blink conditioning paradigms have also implicated both structures in learning, though this is dependent upon the timing parameters of conditioning (Cheng et al., 2008; Christian & Thompson, 2003; Woodruff-Pak, Papka, & Ivry, 1996). In parallel, investigations into cerebello-hippocampal interactions in age-related neurodegenerative disease, particularly AD, stand to be particularly informative. Indeed, implications for this circuit have come from work on eye-blink classical conditioning (Woodruff-Pak et al., 2010; Woodruff-Pak, Papka, Romano, et al., 1996; Woodruff-Pak & Thompson, 1988), as this learning is negatively impacted in AD and in aging. Further focus on this circuit, both with respect to its function in behavior beyond spatial navigation, as well as structural connections between the regions, both cross-sectionally and longitudinally stands to advance our understanding of two key circuits in the aging brain.
To this point, our understanding of cerebello-hippocampal interactions in the human brain is an emerging one. While converging evidence across imaging modalities indicates that there are functionally important interactions between these regions, more direct work focusing on this circuit in learning and memory, and cognition more broadly, is needed. Perhaps most useful however, would be additional detailed work mapping both structural and functional networks linking the two structures. While Arrigo and colleagues (2019) demonstrated in principle that there are white matter linkages in the human brain, the tracing did not use waypoints, the sample size was small, and the direct connection is at odds with subsequent work in rodent models (Watson et al., 2019). A larger-scale follow-up in the human brain to extend this work in the context of the multisynaptic findings in rodents (Watson et al., 2009) stands to build on this initial human work (Arrigo et al., 2014) and provide novel insights into the underlying structural links in this circuit. In parallel, functional connectivity mapping also stands to be useful for further understanding the interrelations between the cerebellum and hippocampus.
While invasive stimulation and recording methods have been especially informative in rodent models, this approach untenable in the human brain (with some exceptions, and these typically involve recording and stimulation from only one site (J. Jacobs et al., 2013; Sunthana et al., 2012)). However, advances in non-invasive brain stimulation methodologies allow for cerebellar stimulation and can be combined with neuroimaging to better investigate the dynamics of the functional interactions between the cerebellum and hippocampus. While this is limited as it is likely to be unidirectional with the cerebellum as the primary stimulation target, further advances in the depth and focality of stimulation methodologies may eventually allow for hippocampal targeting as well. Notably, as outlined in Table 1, many of the interactions between these regions are related to Lobule VI and Crus I of the cerebellum, making these particularly interesting stimulation targets for such an investigation.
The cognitive neuroscience of aging literature has to this point greatly advanced our understanding of brain differences and changes in advanced age, and age-related neurodegenerative disease. However, we still face many challenges, particularly as we consider novel paths to remediation for improved quality of life and function in older adults, and approaches to understanding neurodegenerative disease, particularly Alzheimer’s disease. In recent years, there has been an emergence of exciting work demonstrating a functionally important circuit between the cerebellum and hippocampus, though there is historical evidence for these interactions which provides further support for investigations of this circuit. While the cerebellum is an emerging region of interest in the study of aging and Alzheimer’s disease, and increased focus on this region, its functions, and its networks would be broadly informative, a particular focus on the cerebello-hippocampal circuit stands to be especially useful. Such work may advance our understanding of cognitive and motor deficits in advanced age and may also provide insights into the behavioral patterns seen in Alzheimer’s disease. The bi-directional influences of the cerebellum and hippocampus on one another suggest that these two regions that are important for learning, memory, and spatial navigation, may act in concert more broadly to influence behavior in advanced age.