In a recent paper [A. Alberucci, C. Jisha, N. Smyth, and G. Assanto, Phys. Rev. A 91, 013841 (2015)], Alberucci et al. have studied the propagation of bright spatial solitary waves in highly nonlocal media. We find that the main results in that and related papers, concerning soliton shape and dynamics, based on the accessible soliton (AS) approximation, are incorrect; the correct results have already been published by others. These and other inconsistencies in the paper follow from the problems in applying the AS approximation in earlier papers by the group that propagated to the later papers. The accessible soliton theory cannot describe accurately the features and dynamics of solitons in highly nonlocal media. Snyder and Mitchell introduced in 1997 a model of nonlinearity whose response is highly nonlocal [1]. They proposed an elegant theoretical model, intimately connected with the linear harmonic oscillator that described complex soliton dynamics in simple terms. Because of the simplicity of the theory, they coined the term accessible solitons (ASs) for these optical spatial solitary waves.An early experimental observation of accessible solitons was reported in [2,3]: "We believe that most of the observed spatial optical solitons in nematic liquid crystals (NLCs) are indeed accessible solitons, inasmuch as NLC are highly nonlocal." The same authors have determined the basic beam evolution laws for highly nonlocal NLCs in [2,3], which are later elaborated in [4,5]. Equation (6) in [3] was used to interpret the experiments; the authors claimed good agreement between the data and model predictions.However, straightforward application of the AS approximation, even in nonlinear media with almost infinite range of nonlocality, inevitably leads to additional problems [2,3,6], because there exists no real physical medium without boundaries and without losses. To include the impact of the finite size of the sample, we developed a variational approach (VA) to solitons in nonlinear media with long-range nonlocality, such as NLCs [7]. Our VA results are corroborated by numerical simulations, and even have invited a comment [8,9].We highlighted the differences between accessible soliton approximation and variational approach in nonlocal nonlinear media in [10]. The major differences are linked to the soliton shape and dynamics:where R is the beam width and Λ is the period of small oscillations around the equilibrium. Thus, in crucial characteristics, the AS approximation is an oversimplification that at best can only qualitatively apply to solitons in highly nonlocal media. The first published accurate quantitative correction to AS approximation was presented in [7], direct comparison between AS theory and VA was given in [10], while a complete comparison in D=1,2,3 spatial dimensions can be found in [11].In [4], published after our papers [7] and [10], Alberucci et al. discussed the main features of ASs in realistic diffusive self-focusing media. Authors concluded that "the highly nonlocal (AS) model does not accuratel...