I have been asked repeatedly how I add new content to my existing knowledge, so I will disclose the exact imagery I use for hyperlinking.
I use several forms of hyperlinking, and select the form most suitable to the current document.
For structured information that follows accurate logical order I use some sort of combination between mindmap and UML. I create icons for each marker, and build an accurate data structure for storing information. In this case links to other data look like indexes/pointers. On each link I add direction (does not need to be bidirectional), line width (strength of connection), color (black=comment, red=need processing, green=new idea, blue=supporting data) and a small logical operator (=, ?, <>) so I can follow it.
For loci type information [high level visualization] I usually add gateways with pictures above the gateway. The style of gateway is typically dictated by loci.
For classical visual markers (3D objects) I add physical strings connecting to the object in the different world. I add material, color, thickness and texture to the string to remember connections.
I always double-encode links by imagining clickable hyperlink (usually with colored background) in the text I am reading. The link location in the text itself helps, especially if I need to read the text again after some time.
Usually I test the important links by mentally clicking hyperlink and opening a page, pulling a rope and generating response, sending electric signal, jumping through the portal etc.
Very often a link connects different representations and I have some dictionary of how I transform from one universe to another. For example I can jump through portal and transform into electronic signal, or pull a rope and activate some imaginary transporter.
This is about as detailed as I can get without showing specific examples. If you do want me to show how I address specific examples, please do come up with such examples.