A code $C \colon \{0,1\}^k \to \{0,1\}^n$ is a $q$-query locally decodable code ($q$-LDC) if one can recover any chosen bit $b_i$ of the message $b \in \{0,1\}^k$ with good confidence by querying a corrupted string $\tilde{x}$ of the codeword $x = C(b)$ in at most $q$ coordinates. For $2$ queries, the Hadamard code is a $2$-LDC of length $n = 2^k$, and this code is in fact essentially optimal. For $q \geq 3$, there is a large gap in our understanding: the best constructions achieve $n = \exp(k^{o(1)})$, while prior to the recent work of [AGKM23], the best lower bounds were $n \geq \tilde{\Omega}(k^{\frac{q}{q-2}})$ for $q$ even and $n \geq \tilde{\Omega}(k^{\frac{q+1}{q-1}})$ for $q$ odd.

The recent work of [AGKM23] used spectral methods to prove a lower bound of $n \geq \tilde{\Omega}(k^3)$ for $q = 3$, thus achieving the "$k^{\frac{q}{q-2}}$ bound" for an odd value of $q$. However, their proof does not extend to any odd $q \geq 5$. In this paper, we prove a $q$-LDC lower bound of $n \geq \tilde{\Omega}(k^{\frac{q}{q-2}})$ for any odd $q$. Our key technical idea is the use of an imbalanced bipartite Kikuchi graph, which gives a simpler method to analyze spectral refutations of odd arity XOR without using the standard "Cauchy-Schwarz trick", a trick that typically produces random matrices with correlated entries and makes the analysis for odd arity XOR significantly more complicated than even arity XOR.