(Related commissure) - It is often the case in neural networks that signals are carried en-mass from one set of places to another, from many places to a single place, or from one place to many places.

In biological brains the bundles of axons that carry these signals are generally referred to as commissures, though they are occasionally called bundles, or nerve-bundles as well. One of the largest, and most well known examples of a commissure, in biological brains, is the Corpus Callosum, which connects the left and right hemispheres of the brain.


Bundles in Netlab

In Netlab, Bundles are a special construct that are specifiable in the network description language (Noodle™). Bundles in Netlab form a sort of connection-class, which is referred to generally as a Bundle or Axon-Bundle. As stated, in Noodle™ they are referred to specifically as a Bundle. They carry sets of axons which can originate from a variety of sources, such as neurons, and input devices.

Bundles in Netlab are loosely analogous to biological commissures, though they are an abstraction, and therefor more generalized. For example, while Netlab's bundles can carry a bundle of nerves from one location to another in a neural network, they can also be connected directly to the inputs of single neurons. In this capacity, while they still represent a bundle of nerves, they may also be thought of as extending the neuron's synaptic input-space, and therefor serve (again, loosely) as a dendrite of the neuron to which they are attached.

Netlab's bundles are also given a notational convention for the sake of schematically diagramming networks.

Schematic notation for bundles

Spot notation: 
        <BundleName> 

Bus notation: 
        ############### 

Combined: 
        #####<BundleName> 
                 # 
                 # 
                 # 

Connecting to a bundle 
                 __ 
                |  |       /--------<CorpCol> 
     <xBndl>####|  |       |           
                |   >------+ 
       ==+======|  |       | 
         #      |__|       | 
         #               ##+###### 
         # 

In schematic-notational context, bundles are closely spaced double lines with cross hatch patterns (railroad tracks). Color of a bundle (if color is used) must be different from axons, and neurons. The crosshatch lines are not required to overhang the long lines.