docs: add summary and conclusions

docs/final-report/conclusion.Rmd

@@ -0,0 +1,5 @@
+Muscle movement is the result of coordination between the body’s nervous and muscular system. To move a muscle, the brain releases action potential, a short burst of electrical activity. Action potential propagates via the nervous system, triggering a motor unit. The motor unit generates a specific type of action potential, known as motor unit action potential (MUAP), which causes the attached muscle fibers to contract. 
+
+A technique known as electromyography (EMG), uses electrodes to measure the electrical activity across a single muscle. The raw signal is the result of many motor unit action potential peaks constructively and destructively interfering with each other.  A blind source separation algorithm, as described by @negro_muceli_castronovo_holobar_farina_2016, can decompose raw EMG signals into several individual electrical signals that can be ascribed to singular muscle units.
+
+Our final data product presented is an open-source Python package, `EMGdecomPy`, containing two main elements. The first component is the reconstructed decomposition algorithm as described in the pseudocode of [@negro_muceli_castronovo_holobar_farina_2016]. The algorithm has been slightly adjusted and validated using data from [@Hug2021]. The second component is a Jupyter notebook that provides a template for users to interactively view the decomposed output of the algorithm. 

docs/final-report/summary.Rmd

@@ -0,0 +1,4 @@
+The decomposition of EMG signals into individual motor units is a critical technique for understanding the relationship between the nervous system and the muscular system. The Sensorimotor Physiology Laboratory’s research currently decomposes using a free software from OT Bioelettronica, whose code is based on a paper from @negro_muceli_castronovo_holobar_farina_2016. As this software has several shortcomings, over the course of capstone, we have replicated and tested the decomposition algorithm into an open-source Python package called `EMGdecomPy`. Additionally, we have included a tool to visualize the decomposed output of the algorithm. 
+
+`EMGdecomPy` satisfies the needs of the Sensorimotor Physiology Lab’s research. The decomposition algorithm was adapted from the pseudocode as described by @negro_muceli_castronovo_holobar_farina_2016. The phases of decomposition (pre-processing, separation, and refinement, visualization) were broken up into individual functions. These functions were individually tested and extensively documented for the sake of transparency. Finally, per the partner’s request, the 100 second time limit was removed from the software. 
+While the sequential steps to decompose raw EMG signals were given, there was room for interpretation and improvement in the implementation of each step. As `EMGdecomPy`’s open-source licensing encourages other parties to expand on the package, future recommendations for improvement include improving code speed through optimization,  improving decomposition accuracy through domain knowledge, as well as implementing the aforementioned algorithm relearning feature.