Notes for week 5

Barbara

Shimon

The functional need

• Language: a corpus as a graph (Edelman et al., 2003; 2005).


Synfire chains


• Representing words/concepts by sets of neurons (e.g., Spivey, 2006).
• Representing word/concept sequences by...? Synfire chains! (e.g., Pulvermüller, 1999).

Abeles

• The basic idea: 1982.
• Corticonics: Neural Circuits of the. Cerebral Cortex (1991).

Ikegaya et al.

• "How can neural activity propagate through cortical networks built with weak, stochastic synapses? We find precise repetitions of spontaneous patterns of synaptic inputs in neocortical neurons in vivo and in vitro. These patterns repeat after minutes, maintaining millisecond accuracy. Calcium imaging of slices reveals reactivation of sequences of cells during the occurrence of repeated intracellular synaptic patterns. The spontaneous activity drifts with time, engaging different cells. Sequences of active neurons have distinct spatial structures and are repeated in the same order over tens of seconds, revealing modular temporal dynamics. Higher order sequences are replayed with compressed timing."
• A description of functional multineuron calcium imaging (fMCI; see below) is available at Ikegaya's website.

  

Jun and Jin

• Development of neural circuitry for precise temporal sequences through spontaneous activity, axon remodeling, and synaptic plasticity, J. K. Jun and D. Z. Jin, PLoS One 8:e723 (2007).
• "We show through computer simulation that long synfire chains can develop through spike-time dependent synaptic plasticity [STDP] and axon remodeling --- the pruning of prolific weak connections that follows the emergence of a finite number of strong connections. The formation process begins with a random network. A subset of neurons, called training neurons, intermittently receive superthreshold external input. Gradually, a synfire chain emerges through a recruiting process, in which neurons within the network connect to the tail of the chain started by the training neurons. The model is robust to varying parameters, as well as natural events like neuronal turnover and massive lesions."
• Fig.2:
  • Note that they start with a fully but weakly connected network.
  • "[...] Axon remodeling, in which weak connections from a neuron are pruned once a finite number of super-connections from the same neuron are formed."
• Fig.3: show movie.
  • "Around trial number 180000, the size of the developed network shown in Figure 3 plateaus at 67 groups. At that point, approximately half of the neurons have been recruited, while the other half remain in the pool."
• Note anything strange about Jun & Jin's model?

Polychronization

Izhikevich

• Start with a sparse randomly connected network with STDP and variable axonal conduction delays.
• "From a mathematical point of view, a system with delays is not finite- but infinite-dimensional [...]"
• "We simulated an anatomically realistic model consisting of 100,000 cortical spiking neurons having receptors with AMPA, NMDA, GABA_A, and GABA_B kinetics and long-term and short-term synaptic plasticity."
• Spike-timing-dependent plasticity (STDP).
• "The plasticity takes an initially unstructured network and selects firing patterns that are consistent with the underlying anatomy, thereby creating many strongly connected subgraphs corresponding to polychronous groups. Since STDP is always "ON" in the network, groups constantly appear and disappear; their total number fluctuates between 5000 and 6000. We found a core of 471 groups that appeared and survived the entire duration of 24 hour simulation."
• "Considering toy examples like this one, it would not be surprising that a network of 10¹¹ neurons (which is the size of the human brain) would have more groups than the number of particles in the universe."
• "Figure 17: Time-locked activation of groups representing various features of a stimulus results in binding of the features and increased gamma rhythm. Each group contributes small gamma oscillation to the network gamma. (Left) The oscillations average out during the asynchronous activation. (Right) The oscillations add up during the time-locked activation. Dotted lines: weak reentrant connections between the groups that synchronize (or polychronize) their activation (the groups and the associated gamma rhythm are drawn by hand)."