I am unfamiliar with the term "Predictive Reverse Engineering" but your question seemed very interesting. Informally speaking, "Predictive Reverse Engineering" seems to denote trying to reverse engineer the brain's architecture in a way that is also able to predict brain function, which was not used in the process of constructing the brain's map. In other words, to try to construct a model of the brain from sparse data, which is able to predict the results of novel observations. Like I said, I'm no expert in the subject so I poked around a bit and here is what I found. I hope it helps.
The human brain project (HBP) has a list of key results and publications list. The most noteworthy of all seems to be Markram et al (2015), which is a huge paper that shows the results of a first draft reconstruction of the rat's somatosensory cortex. This paper has a very long suppplementary information section describing how they reconstructed neuronal morphologies from the experimental data.
The introduction of the above article cites many works that communicate how the anatomical and physiological properties of synaptic connections can be characterized (Cobb et al., 1997; Feldmeyer et al., 1999; Frick et al., 2008; Gupta et al., 2000; Mason et al., 1991;
Reyes et al., 1998; Thomson et al., 1993). It also lists some methods used to map pre and postsynaptic neurons for individual neurons, such as: retrograde and anterograde tracers and trans-synaptic viral vectors, imaging with array tomography, and saturated reconstruction with electron microscopy. The references they cite for these methods are: Boyd and Matsubara, 1991; Callaway, 2008; Glenn et al., 1982; Kasthuri et al., 2015; Killackey et al., 1983; Micheva and Smith, 2007; Micheva et al., 2010; Wickersham et al., 2007.
The methods listed above by no way map the entire somatosensory cortex. They gather data from a sparse locations and Markram et al present a computational method to reconstruct the entire microcircuitry, i.e. reverse engineer the somatosensory cortex from sparse observations. They show that they are able to cross validate their algorithmic approach with the data and also account for new experimental observations.
Another paper that you could find interesting is Frackowiak and Markram (2015), "The future of human cerebral cartography: a novel approach", which is a perspectives article on how the brain can be cartographically mapped to understand function.