%0 Journal Article %A Florian Fiebig %A Pawel Herman %A Anders Lansner %T An Indexing Theory for Working Memory based on Fast Hebbian Plasticity %D 2020 %R 10.1523/ENEURO.0374-19.2020 %J eneuro %P ENEURO.0374-19.2020 %X Working memory (WM) is a key component of human memory and cognition. Computational models have been used to study the underlying neural mechanisms, but neglected the important role of short- and long-term memory interactions (STM, LTM) for WM. Here, we investigate these using a novel multi-area spiking neural network model of prefrontal cortex (PFC) and two parieto-temporal cortical areas based on macaque data. We propose a WM indexing theory that explains how PFC could associate, maintain and update multi-modal LTM representations. Our simulations demonstrate how simultaneous, brief multi-modal memory cues could build a temporary joint memory representation as an “index” in PFC by means of fast Hebbian synaptic plasticity. This index can then reactivate spontaneously and thereby also the associated LTM representations. Cueing one LTM item rapidly pattern-completes the associated un-cued item via PFC. The PFC-STM network updates flexibly as new stimuli arrive thereby gradually over-writing older representations.Significance Statement Most if not all computational working memory (WM) models have focused on short-term memory (STM) aspects. However, from the cognitive perspective the interaction of STM with long-term memory (LTM) bears particular relevance since the WM activated LTM representations are considered central to flexible cognition. Here we present a large-scale biologically detailed spiking neural network model accounting for three connected cortical areas to study dynamic STM-LTM interactions that reflect the underlying theoretical concept of memory indexing, adapted to support distributed cortical WM. Our cortex model is constrained by relevant experimental data about cortical neurons, synapses, modularity, and connectivity. It demonstrates encoding, maintenance and flexible updating of multiple items in WM as no single model has done before. It thereby bridges microscopic synaptic effects with macroscopic memory dynamics, and reproduces several key neural phenomena reported in WM experiments. %U https://www.eneuro.org/content/eneuro/early/2020/02/28/ENEURO.0374-19.2020.full.pdf