'Mini analysis' is an unreliable reporter of synaptic changes

Abstract

Analysis of miniature postsynaptic currents (mPSCs, or minis) is one of the most extensively employed approaches to determine the functional properties of synaptic connections. The popularity of this technique stems from its simplicity. Patch-clamp recording is used to observe spontaneous transmission occurring at synapses throughout a neurons dendritic tree. These events are analysed in an attempt to determine quantal synaptic parameters and changes during experimental manipulation. For decades, changes in the amplitude or frequency of mPSCs have been interpreted as evidence for specific synaptic modifications, including changes in presynaptic release sites, postsynaptic receptor abundance or even nanoscale changes in the alignment of synaptic machinery. However, at the majority of brain synapses, these events are small, with many undetectable due to recording noise. Here, using both simulated and experimental data, we demonstrate that interpreting synaptic changes from mPSC datasets is fundamentally fallible. Due to incomplete detection of event distributions, seemingly specific changes in mPSC amplitude or frequency falsely report actual synaptic changes. In addition, probabilistic detection of small events gives false confidence in the completeness of detection and inaccurate determination of quantal size. We not only demonstrate the dense pitfalls of mini analysis, but also establish a method for experimental detection of the detection limit, allowing more robust data analysis and scientific interpretation.