Neural codes and the factor complex (Q2300577): Difference between revisions

Given the set \([n]:=\{1,\dots,n\}\), we call neural code on \(n\) neurons, any subset of \(2^{[n]}\). It is important for these codes to be convex and a class of convex codes is that of \textit{max-intersection-complete codes}. In the referenced paper [\textit{C. Curto} et al., J. Pure Appl. Algebra 223, No. 9, 3919--3940 (2019; Zbl 1415.92044)], the authors ask an algebraic signature for such codes and this is the question addressed by the present paper. Consider \(R=\mathbb{F}_2[x_1,\dots,x_n] \) and define \textit{pseudomonomials} as polynomials of the form \[ \prod_{i \in \sigma} x_i \prod_{j \in \tau} (1-x_j),\quad \tau,\sigma \in 2^{[n]},\, \sigma \cap \tau =\emptyset. \] The ideals generated by such pseudomonomials are called pseudomonomial ideals. For an element of \(c \in 2^{[n]}\), its indicator polynomial is \(\prod_{i \in c} x_i \prod_{j \in [n]\setminus c} (1-x_j)\). Now if we take as generators for an ideal all the indicator polynomials of the non-codewords of a neural code \(C\) we get a neural ideal. For such kind of ideal, say \(J\), we can consider all pseudomonomials that are minimal w.r.t. divisibility with all the pseudomonomials in \(J\) and the set of all those monomials is the canonical form of \(J\), denoted by \(CF(J).\) To a neural code we can also associate a simplicial complex, that is denoted by \(\Delta(C)\) and it is by definition the smallest simplicial complex containing the given code. We denote by \(I(\Delta(C))\) the associated Steiner-Reisner ideal. Now, we can state the result which answer to the question that is posed in [\textit{C. Curto} et al., J. Pure Appl. Algebra 223, No. 9, 3919--3940 (2019; Zbl 1415.92044)]: Theorem 1. A code \(C\) on \(n\) neurons is max-intersection-complete if and only if for every non-monomial \(\phi\) in the canonical form of the neural ideal of \(C\), there exists \(i \in [n]\) such that: \begin{enumerate} \item every associated prime of \(I(\Delta(C))\) that contains \(x_i\) also contains \(\phi\), and \item \( (1 - x_i ) \mid \phi\). \end{enumerate} This theorem is proved using and actually introducing the factor complex of \(C\), another simplicial complex, particularly well-suited to represent the combinatorial properties of \(C\). From Theorem 1, an algorithm is deduced and it is shown to be advantageous with respect to those in literature. The paper also characterize those codes that are closed under intersection of codewords, called intersection-complete codes.