Alzheimer's disease (AD) patients suffer from an inexorable loss of cognitive function due in part to the degeneration of the basal forebrain cholinergic projection neurons. Restoring signalling through the post-synaptic muscarinic acetylcholine receptors that mediate cholinergic signalling is thus a promising therapeutic strategy for the symptomatic management of AD. Moreover muscarinic receptor agonists reduce Aβ42 in animal models, suggesting that this class of compounds could address the underlying disease pathology. Unfortunately, multiple non-selective muscarinic receptor agonists have been tested in AD patients and were not tolerated due to unwanted peripheral cholinergic stimulation. Improved selectivity for M1, the most prominently expressed post-synaptic muscarinic receptor in the cortex and hippocampus, could address these tolerability issues, however the high conservation of the acetylcholine binding site between receptor subtypes has precluded the discovery of selective agonists. To achieve selectivity for M1 by discovering and exploiting an allosteric ligand binding site. We screened a chemical library for positive allosteric modulators of the M1 receptor and assessed chemical leads for selectivity in vitro and efficacy in rodent AD models in vivo. We identified Benzyl Quinolone Carboxylic Acid (BQCA), a small molecule potentiator selective for M1. In CHO cells expressing recombinant human receptor, BQCA sensitizes M1 to acetylcholine 83.9-fold (inflection point = 449 nM), while having no potentiation, agonist, or antagonist effect on M2, M3, or M4 receptors up to 100 μM. In the mouse contextual fear conditioning model of episodic-like memory, BQCA fully reversed the cognitive impairment caused by the non-selective muscarinic antagonist scopolamine, demonstrating a critical role for M1 in this type of memory. Importantly BQCA does not show signs of unwanted peripheral cholinergic stimulation at doses that produce central physiological responses. In vivo BQCA and other muscarinic agonists can specifically reduce cortical Aβ42 in some experiments. However, this effect is likely complex and indirect, since it does not occur in cultured neurons and is not consistently observed across species and disease models. BQCA is a highly selective pharmacological reagent for understanding the normal physiology of the M1 receptor and its potential as a therapeutic target for AD.