Sympathetic ganglia, along with the skeletal neuromuscular junction, have been fundamental in forming our understanding of synapses and neurotransmitter release. From the earliest descriptions of their neuroanatomy and the pioneering work of Eccles (Eccles, 1943) these peripheral ‘little brains’ have given us powerful, insights into how the nervous system operates. Their easy access, relative durability in vitro compared to ‘fickle’ central neurons coupled with their robust synaptic responses made them a good model system for early electrophysiologists. Whilst providing considerable information about how neurons work, the role of the ganglionic synapse in vivo remained elusive. Early on it was recognised (in vitro) that under some circumstances separate subthreshold preganglionic synaptic inputs could summate to evoke action potentials in postganglionic neurons, and it was assumed that integration would play a major role in normal physiology. Landmark experiments involving microneurography in humans and extracellular recordings from animals revealed that the average action potential discharge of pre- and postganglionic sympathetic axons is extremely low <1 Hz average with occasional high frequency bursts of several action potentials (reviewed in Janig, 2006). Preceding these experiments, in vitro data using sharp microelectrodes revealed that the typical profile of a paravertebral sympathetic neuron was one very strong input with a few small subthreshold (weak) inputs. Simple modelling of the time course and amplitude of weak excitatory junction potentials based on membrane time constants derived from the best sharp electrode recordings led to the conclusion that summation of secondary subthreshold synaptic inputs were unlikely to significantly amplify central sympathetic outflow. A small number of technically heroic in vivo recording studies essentially confirmed little role for amplification through convergent secondary inputs in paravertebral ganglia (McLachlan, 2003). A major feature of the sympathetic nervous system is the wide divergence where each preganglionic neuron innervates a number of postganglionic neurons (Janig, 2006). Collectively the above data left us with the concept that the dominant role of paravertebral sympathetic ganglia was to distribute central command from a small number of preganglionic neurons to a large peripheral territory. This begged the question, why maintain functional convergence? In the 1980s and 90s the development of the whole cell patch clamp recording technique revolutionised the study of central synapses by allowed high fidelity recordings from CNS neurons in vitro and later in vivo. Unfortunately it turned out that the properties that made sympathetic ganglia so resilient in vitro made them extremely difficult to patch clamp in vitro, and (so far) impossible in vivo. Their connective tissue sheath and the fact that each neuron is carefully gift wrapped by several Schwann cells left no neuronal membrane to patch onto. Whole cell patch clamp recording became reserved for studies on dissociated sympathetic neurons and sharp intracellular microelectrodes reserved for studying the sympathetic ganglionic synapse. Early on it was recognised that there was a profound difference in electrophysiological data obtained using the whole cell patch clamp versus sharp microelectrodes. Input resistance and membrane time constant were an order of magnitude higher using patch clamp (Gola & Niel, 1993). The paper by Springer, Kullmann and Horn in this issue of The Journal of Physiology (Springer et al. 2015) revisits the question, under what conditions do sympathetic neurons integrate synaptic inputs? Advances in computational speed in the last decade has allowed investigators to evoke ‘virtual synapses’ via dynamic clamp software. In addition dynamic clamp allowed a ‘virtual leak conductance’ to be applied, simulating those introduced by sharp microelectrodes. Importantly, data from dissociated neurons were validated in intact ganglia that retained their dendritic arbor. The main outcome was that using patterns of ‘virtual’ preganglionic activity taken from in vivo studies, sympathetic neurons amplified preganglionic discharge ∼2.5-fold. This amplification was hidden when a virtual leak conductance was used to mimic intracellular sharp microelectrode recordings. Another finding was the profound differences in action potential discharge properties found in intact and dissociated neurons under whole cell patch clamp configuration compared with published data from intracellular electrodes. Many more neurons were slowly adapting compared with data from sharp recordings. If sympathetic ganglia do behave as activity-dependent amplifiers of central outflow, a plethora of questions arise, or are re-opened. We need to rethink the size of the functional ‘autonomic neural unit’ as defined by Purves (reviewed in Janig, 2006). It may be larger than previously thought and more plastic. Autonomic ganglia sit in a ‘soup’ of neuromodulators from peptidergic co-transmitters to circulating hormones and locally released cytokines. Previous estimates of their likely impact in vivo were based on recordings that underestimated membrane time constant and input resistance. Consequently much lower concentrations of these neuromodulators may be more effective than previously thought. A great deal of effort has gone into defining sympathetic neurons based on their responses to depolarizing current injection, and the importance of the selective expression of voltage-dependent currents such as IM, IA and IH. Synaptic gain was only weakly determined by action potential discharge phenotype. This begs the question, why do sympathetic neurons show substantial differences in the expression of voltage-dependent currents? So far dynamic clamp has only been applied to paravertebral sympathetic neurons as have all in vivo sharp recordings to date. Motility-regulating sympathetic neurons in prevertebral ganglia have substantially different reflex activation to cardiovascular neurons. Furthermore these neurons receive many subthreshold inputs from preganglionic and intestinofugal neurons. The application of realistic patterns of activity to these neurons should provide insight into the control of viscera. Finally, what are the clinical implications? In a subset of patients, hypertension is likely to be maintained by increased sympathetic outflow (Esler, 2014). Perhaps the ganglionic synapse is back on the table as a potential therapeutic target?