Revisiting the Reported Period of FRB 20201124A Using MCMC Methods

TL;DR

The paper tackles whether repeating FRBs exhibit true periodic modulation compatible with magnetar spin by introducing a fast phase-folding approach combined with Markov Chain Monte Carlo (MCMC) estimation. The method models the folded phase as a Von Mises distribution with parameters $(\mu,\kappa)$ and jointly estimates the period $P$ along with phase and concentration parameters using the emcee sampler, enforcing a resolution condition $\delta P \cdot T / P^{2} \le 0.1$. Data from FAST observations of FRB 20201124A (MJD 59307–59360) are partitioned into daily segments and analyzed with 25 walkers and $10^{5}$ steps per segment, revealing major $P$-peaks near the reported $\sim 1.7$ s in some days but no consistent period across all segments. The results validate the method as a rapid coarse search useful for exploring magnetar-spin modulation in FRBs, while also highlighting the importance of cross-segment consistency and the current lack of a universal periodicity in this source.

Abstract

Fast radio bursts (FRBs) are millisecond-duration radio transients whose physical origin remains uncertain. Magnetar-based models, motivated by observed properties such as polarization and large rotation measures, suggest that FRB emission may be modulated by the magnetar spin period. We present an efficient method to search for periodic signals in repeating FRBs by combining phase folding and Markov Chain Monte Carlo (MCMC) parameter estimation. Our method accelerates period searches. We test the method using observational data from repeater FRB 20201124A, and show that it can recover reported candidate periods.

Figures (4)

• Figure 1: Schematic diagram of a Von Mises distribution. The simulated data represent a concentrated phase profile, which can be well described by a Von Mises distribution. Therefore, we adopt the Von Mises distribution to model the optimal profile.
• Figure 2: (a) Parameter estimation was performed based on FAST observations on MJD 59310. The primary peak of the period $P$ occurs at 1.680344–1.706105 s ($1\sigma$). The parameter $\phi_0$ exhibits a secondary peak around 0, which may correspond to a minor peak in $P$. (b) Data segment from MJD 59347. Several minor peaks of $P$ are present, which we attribute to the short duration of the data causes a multiple peaks. The major peak lies at 1.707917–1.707957 s ($1\sigma$). These two period search results are consistent with 2025arXiv250312013D.
• Figure 3: data segment: (a) MJD 59309, (b) MJD 59311, (c) MJD 59318, (d) MJD 59329. (e) MJD 59337, (f) MJD 59339. This figure shows representative MCMC results from our study. Certain corner plots, shown in (a), (c), and (f), exhibit good convergence. In contrast, the corner plot in (b) appears highly chaotic, possibly due to either a lack of true periodicity in the data or the short data segment, which can easily produce spurious period detections.
• Figure 4: The normalized plot of the period search over (0,10) s shows no evidence of a preferred period. The dotted line indicates the location of the maximum count, which does not reveal any significant or special value.