Recordings of electrical signatures of the brain, via electroencephalography (EEG), magnetoencephalography (MEG), electrocorticography (ECoG), and local field potential (LFP), have consistently reported rhythmic patterns in neural activity. Brain rhythms, also known as oscillations, are linked with numerous cognitive functions and group into several oscillatory bands (Buzsáki, 2006). Beta oscillations (∼13–30 Hz) are commonly implicated in sensorimotor processing (Pfurtscheller and Lopes da Silva, 1999; Baker, 2007). These oscillations are established during stable postures and are decreased during active states, such as movement planning and execution (Engel and Fries, 2010; Kilavik et al., 2013). A decrease in the amplitude of beta oscillations across sensorimotor areas is seen just prior to and during movement execution. Conversely, an increase of beta amplitude above baseline levels is observed following movement cessation. We refer to these two phenomena as movement related beta decrease (MRBD) and post-movement beta rebound (PMBR), respectively (Figure 1A). The two principal sources of beta are sensorimotor cortex (Jensen et al., 2005; Roopun et al., 2006; Kramer et al., 2008; Yamawaki et al., 2008; Kopell et al., 2011) and basal ganglia (Figure 1C; Holgado et al., 2010; McCarthy et al., 2011; Tachibana et al., 2011; Mirzaei et al., 2017). There is debate as to whether they originate independently in each area or if they are an emergent property of the cortico-basal ganglia networks (Pavlides et al., 2015; Sherman et al., 2016; Reis et al., 2019).

The scope of this review is to report the most recent interpretations of the role of beta oscillations in the sensorimotor system. We will do so by giving details of how beta oscillations are affected by healthy aging (Rossiter et al., 2014b) and motor learning (Tan et al., 2014, 2016; Torrecillos et al., 2015; Haar and Faisal, 2020) as well as how they are altered in various conditions, most notably Parkinson’s disease (PD) and stroke, characterized by severe motor symptoms (Brown, 2006; Rossiter et al., 2014a). We will also cover how beta oscillations can be manipulated experimentally using drugs and brain stimulation techniques. We will conclude by discussing how assessing the transient nature of beta oscillations could enrich our understating of its functional role (Feingold et al., 2015; Sherman et al., 2016; Shin et al., 2017).

Although several studies have extensively investigated beta oscillatory dynamics (Table 1) [for review, see Kilavik et al. (2013)], its functional role in the sensorimotor processing is still not fully understood. One source of interest lies in beta’s tendency to fluctuate during movement. MRBD is present during spontaneous and triggered movements (Kilavik et al., 2013), while successful movement cancelation is associated with an increase in beta (Swann et al., 2009, 2012; Wagner et al., 2018; Jana et al., 2020). MRBD also occurs when no muscle contraction is required (i.e., motor imagery or action observation) and is rather insensitive to parameters like movement type or effector (Miller et al., 2010; Kilavik et al., 2013). These observations led Engel and Fries (2010) to propose a role for beta as an active process, which interferes with the encoding of incoming information while promoting the existing state—i.e., “status-quo”—of the system. Therefore, instead of being a proxy for the level of activity of the sensorimotor network, beta oscillations act as a top-down inhibitory rhythm during motor and cognitive tasks. Following the “status-quo” hypothesis, MRBD and PMBR could be interpreted as endogenous fluctuations of beta level during a motor set, with the former necessary for releasing the inhibition and allowing the initiation of a motor plan, while the latter preserves the existing motor states from internal and external sources of noise.

Post-movement beta rebound has also been interpreted more specifically as an indicator of movement outcome processing (Baker, 2007). Supporting evidence stems from findings showing PMBR is modulated by passive movements (Cassim et al., 2001; Alegre et al., 2002) and by kinematic errors (Tan et al., 2014). More recently, Tan et al. (2016) reported that the level of PMBR over the sensorimotor cortex serves as an index of confidence in the prediction of a motor outcome, also known as the forward model.

Together, these studies propose a role for sensorimotor beta oscillations encompassing multiple functions. We suggest several reasons why interpretations of beta’s role are so varied.

First, beta involves several types of oscillations in distinct frequency bands (Kopell et al., 2011). PD studies show that oscillatory activity through the cortico-basal ganglia network is segregated into low (14–20 Hz) and high beta frequencies (>24 Hz) both in humans (López-Azcárate et al., 2010; Litvak et al., 2011) and rats (West et al., 2018). Furthermore, dopamine levels affect low and high beta rhythms differently (Brown et al., 2001; Priori et al., 2004; Marceglia et al., 2006). One theory proposes that low-beta has an “anti-kinetic” role (Brown, 2003; Chandrasekaran et al., 2019), while high-beta reflects attention and sensory cue anticipation (Saleh et al., 2010; Kilavik et al., 2014; Chandrasekaran et al., 2019).

Secondly, some studies have observed that MRBD and PMBR have a different spatial distribution (with MRBD localizing to postcentral gyrus and PMBR to precentral gyrus) and could represent independent events (Figure 1B; Jurkiewicz et al., 2006; Miller et al., 2007; Alegre et al., 2008; Gaetz et al., 2011; Muthukumaraswamy et al., 2013).

Finally, recent studies suggest that beta shows a burst-like activity, rather than being sustained over time. Therefore, theories of its functional role should integrate this transient nature and the potential implications for behavior (see section “Beta Oscillations as a Transient Rhythm”).

In previous reports, beta oscillations were typically considered as a repeated cycle of oscillatory activity sustained over time. A growing number of studies, however, are reconsidering beta oscillations as brief bursts of temporally localized activity (Feingold et al., 2015; Sherman et al., 2016; Shin et al., 2017; Tinkhauser et al., 2017; Torrecillos et al., 2018; Little et al., 2019). The idea of oscillations sustained across time stems from the standard analysis procedure of averaging oscillatory activity across many repeated trials (Jones, 2016). Sherman et al. (2016) analyzed source-localized human MEG data looking at spontaneous activity during rest in primary somatosensory cortex. In single trials, beta emerged transiently as a sudden increase in power typically lasting < 150 ms, but when the same data was averaged over many trials, continuous oscillations appeared in the time-frequency spectrogram. Feingold et al. (2015) showed how beta-band events in LFPs occurred predominantly in brief bursts both in the motor-premotor cortex and in the striatum of monkeys performing self-timed movement tasks. Single trial analysis revealed that variations in averaged oscillatory power were expressed by variations in burst density of beta events. Thus, the authors suggest, beta synchronization and desynchronization reflect the probability of occurrence of a brief bursting event, rather than representing modulation of the strength of a sustained oscillation. Finally, Shin et al. (2017) showed that the rate of transient pre-stimulus beta events in the primary somatosensory cortex was the most consistent predictor of stimulus detection in humans and mice, while also being strongly correlated with average beta power. Taken together, cumulative evidence across brain areas, recording modalities and species support the notion of beta as a transient event. Trial averaged beta may conceal the functional importance of the underlying bursting activity, offering a less detailed interpretation of beta’s role in the sensorimotor system. Whilst the averaged amplitude of beta power does seem to correlate with rate of beta events, implying that this new analysis may not contradict what has gone before, it will allow for more detailed exploration which may in turn help to consolidate the details of beta’s role in sensorimotor integration.

Many of the studies in this review link to the idea that beta oscillations at rest ensure stability and the “status quo” of the motor system. Differently, studies focusing on PMBR highlight a much more nuanced relationship with movement, potentially encoding features of motor learning and error-salience. Although beta’s role is still not fully elucidated, it is unlikely to reflect pure sensory or motor processes. Instead, beta has become more broadly implicated in endogenous top-down processing and sensorimotor integration.

Future studies should work toward establishing a causal role of beta in the sensorimotor system. We described two distinct methods, pharmacology and brain stimulation, that can be effectively used to manipulate beta and infer causality. Furthermore, studies with patient groups with selective beta impairments, such as PD and stroke, can provide invaluable evidence of the physiological relevance of beta in motor control and motor learning. What should also be at the forefront in our minds is the richness of detail that may be gleaned with single trial analysis of beta events, which could return additional features to characterize and understand beta’s role in the sensorimotor system.

Both authors contributed to the conception and design of the review, to the interpretation of the relevant literature, and to the writing and editing of the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.