Here's a way to consider the differences between Ne and Ni.
In meterology, there exist thousands of weather stations around the country picking up the wind speed and direction at each station. The data collected is then drawn on a map and a "bulb" is used at the location of each weather station on the map to draw the direction the wind is blowing from.
The bulbs look like this:
The "bulbs" for each station then are drawn on a map like this:
The lines coming out of each circle (indicating a weather station) point in the direction the wind is blowing at each weather station.
Once you have all the weather station data points with the wind lines pointing in their respective directions, you can begin to see the overall wind/frontal patterns that exist around the country. You also can pinpoint a rough location of where a center of high pressure and a center of low pressure exist, since wind is caused by a steep pressure gradient (noted by many blue isobars close together on the map), a drastic change in pressure over a small area.
- - -
I see Ne as those weather station "bulbs" giving an isolated "reading" of how the internal dynamic can be approximated at each local context. I see Ni as the overall atmospheric pressure system swirling/fluctuating within the atmosphere that is behind all the changes in wind patterns detected at each locale.
The idea behind this is that I notice that Ne tends to run around things in the "N-domain", in a way "pointing to the truth" from different vantage points, like with the weather station map bulbs. Ni feels the trend/dynamic directly which Ne people isolate, "approximate", and permute in different contexts. Ne is like a series of mirrors that are in alignment reflecting off each other.
With this example, the Ne perspective sees the weather system as an outside observer looking at the map (essentially from the perspective we're seeing when looking at the map). It sees all the different weather station points, takes into account the Si context at each specific point (thundery skies, temperature, topography, etc), and then one of the introverted judging functions (Ti or Fi) will come in and model the data collected at the different points into something coherent. Ti will model it into an explicit structure, Fi implicit. Ti or Fi will take the multiple Ne data station points and say "Over here we can label an area of high pressure, and we can draw isobars close together to illustrate a steep pressure gradient (as shown over the upper midwest in the map I pasted)".
The Ni perspective is from the point of view of the actual pressure field itself. It perceives the changes in gradient and the unified wind pattern as a part of itself and directly experiences that essence that results in any ostensible changes observed.
Perhaps this weather metaphor could be used to model Se and Si.
What do you think? Feel free to clarify the metaphor and tie in other functions as to how they'd play into this metaphor; I'm still fleshing it out in my mind.