Brain Rhythms Out of Step: Beta Waves and Neurons Disagree in Parkinson’s Region

The electrical rhythms of the brain have long been treated as faithful messengers of what neurons are actually doing. When researchers record a strong oscillation in a brain region, the working assumption has often been that the underlying cells are firing in step with that rhythm, their electrical pulses riding the crest of each wave. A new study published in NPJ Parkinson’s Disease throws a wrench into that comfortable assumption for one of the most clinically important targets in Parkinson’s disease: the subthalamic nucleus. There, the researchers found that beta-band activity measured in the local field potential, the aggregated electrical signal that clinicians and scientists rely on heavily, does not consistently line up with the burst firing of individual neurons. In some cases the bursts occur in phase with the beta rhythm, and in others they occur in antiphase, and the relationship can be inconsistent within the same recording.

The subthalamic nucleus, a small lens-shaped structure deep in the brain, sits at a critical junction in the basal ganglia circuitry that governs movement. In Parkinson’s disease, as dopamine-producing neurons in the substantia nigra degenerate, this circuitry falls into disarray. One of the most robust electrophysiological signatures of the disease is an exaggeration of beta-band oscillations, rhythmic fluctuations in the roughly 13 to 30 hertz range, in the subthalamic nucleus. These oscillations have been correlated with the cardinal motor symptoms of Parkinson’s, including bradykinesia, rigidity, and tremor, and their suppression through deep brain stimulation is associated with therapeutic benefit. This has made beta activity a cornerstone biomarker, both for understanding the disease and for engineering next-generation adaptive stimulation devices that deliver therapy only when the pathological signal is detected.

Yet the meaning of the local field potential remains one of the enduring puzzles of systems neuroscience. The LFP is thought to arise primarily from summed synaptic currents flowing across populations of neurons, filtered by the passive electrical properties of the surrounding tissue. Spikes, by contrast, are the discrete action potentials emitted by individual cells, and they contribute comparatively little to the field potential. When researchers observe beta oscillations in the LFP, they infer that the synaptic inputs to subthalamic neurons are oscillating, and they often further assume that the neurons’ output spikes must therefore be modulated in phase with that rhythm. The new findings complicate that second inference. Burst discharges from subthalamic neurons, the study reports, can be found both in phase and in antiphase with the simultaneously recorded beta LFP, and neither relationship dominates in a stable, predictable way.

This inconsistency matters because the phase of a spike relative to an oscillation is not an incidental detail. In the theoretical framework of neural coding, the timing of action potentials relative to population rhythms carries information and shapes downstream communication. Two neurons that receive the same oscillating input but burst on opposite phases of it are, in a functional sense, responding oppositely: one fires when the population signal peaks, the other when it troughs. If the spiking output of the subthalamic nucleus is split between in-phase and antiphase bursting, then a single LFP beta measurement cannot reliably indicate whether the neurons it governs are firing together or in opposition. The aggregate signal may look identical in both situations while the underlying cellular behavior differs fundamentally.

The implications ripple outward into several active areas of research and clinical development. Adaptive deep brain stimulation, one of the most promising advances in the field, uses real-time measurement of beta-band LFP to trigger or modulate stimulation. Devices currently in clinical trials, and early sensing-capable implants already in patients, treat elevated beta power as a proxy for the pathological state of the circuit. If beta power does not map consistently onto the bursting behavior of subthalamic neurons, then the biomarker may sometimes capture a circuit state that differs from the cellular dynamics the stimulation is intended to disrupt. This does not invalidate the approach, since beta suppression demonstrably correlates with symptom relief, but it does suggest that the chain of inference from LFP measurement to neuronal mechanism is weaker than often assumed.

The findings also speak to long-standing debates about the origins of pathological beta oscillations in Parkinson’s disease. Competing models assign different weights to the subthalamic nucleus itself, to its reciprocal connections with the external segment of the globus pallidus, and to cortical input delivered through the hyperdirect pathway. If subthalamic neurons burst in antiphase with the local field oscillation in a substantial fraction of cases, then the relationship between synaptic drive and spiking output in the nucleus is more heterogeneous than many models allow. Inhibitory input from the globus pallidus, which arrives as rhythmic bursts in parkinsonian conditions, could plausibly produce bursts of spikes in subthalamic neurons that occur during phases of suppressed synaptic depolarization in the surrounding population, contributing a sign-inverted component to the spike-LFP relationship. The new results are consistent with such heterogeneity in the sources and signs of rhythmic drive.

Methodologically, the study underscores the importance of examining spike-field relationships at the level of individual units rather than relying on population averages. When spikes from many neurons are pooled, in-phase and antiphase bursting can partially cancel, yielding a weak or ambiguous phase locking statistic that might be dismissed as noise. The more informative observation is that both relationships coexist, often within the same recording session, which means the averaging itself obscures the underlying structure. This echoes a broader lesson in electrophysiology: aggregate signals such as the LFP, electrocorticogram, and scalp EEG are powerful and clinically practical, but their interpretation requires careful attention to the geometry of the sources and the diversity of the cellular responses they summarize.

For patients and clinicians, the immediate practical consequences are limited but worth stating precisely. The therapeutic effectiveness of deep brain stimulation does not depend on the spike-field relationship being in phase; high-frequency stimulation suppresses symptoms regardless, presumably by driving the circuit into a more regular, information-rich regime that disrupts pathological patterning. The concern is prospective: as the field moves toward closed-loop therapies that decode brain state from field potentials, and toward brain-computer interfaces that treat oscillatory phase as a control signal, the assumption that phase reflects cellular firing must be tested rather than assumed. The new results provide a concrete, clinically relevant example where that assumption fails, at least intermittently, in a structure that is the single most common target of functional neurosurgery.

The research also raises questions that future work will need to address. Whether the inconsistent phase relationships reflect differences among neuron types within the subthalamic nucleus, shifts across behavioral states such as rest and movement, fluctuations in the balance of excitatory and inhibitory drive over time, or artifacts of how recording contacts sample spatially extended oscillatory sources remains to be determined. Longitudinal recordings from implanted patients, combined with computational models of the basal ganglia network, offer a path toward resolving which cellular configurations generate which field signatures. Such work could ultimately refine the biomarkers used in adaptive stimulation, allowing devices to distinguish circuit states that currently look identical in the LFP.

In the broader arc of neuroscience, the study is a reminder that the brain’s rhythms are not monolithic expressions of collective firing but composites whose relationship to cellular activity is contingent and, in the parkinsonian subthalamic nucleus, demonstrably inconsistent. The beta oscillation will remain a valuable clinical signal, and the correlation between its power and Parkinsonian symptoms is not in dispute. What the new findings erode is the simpler narrative in which the rhythm and the spikes move as one. In the subthalamic nucleus, neurons can march with the beta wave or against it, and the field potential alone cannot tell an observer which. For a field betting increasingly on oscillations as the language of pathological brain circuits, that ambiguity is a finding worth taking seriously.

Subject of Research: The relationship between subthalamic local field potential beta oscillations and neuronal burst firing in Parkinson's disease.

Article Title: Inconsistent subthalamic local field potential beta activity amid in- and antiphasic neuronal bursts

Article References: Inconsistent subthalamic local field potential beta activity amid in- and antiphasic neuronal bursts. (n.d.). https://doi.org/10.1038/s41531-026-01531-4

Image Credits: AI Generated

DOI: 10.1038/s41531-026-01531-4

Keywords: Parkinson's disease, subthalamic nucleus, local field potentials, beta oscillations, neuronal bursting, deep brain stimulation, basal ganglia, electrophysiology, neural coding, phase-amplitude coupling, brain-computer interfaces, movement disorders