Add picture and revise the text

Author: neuromechanist

assets/images/studyclust14.png

@@ -309,13 +309,13 @@ Press *Ok*. The cluster editing interface detailed in one of the following secti
 
 Optimal Kmeans clustering
 -----------------
-We have recently added a new feature to the *kmeans* clustering algorithm in EEGLAB. This feature allows you to find the optimal number of clusters for your data. To use this feature, you need to have the [MATLAB Statistics and Machine Learning Toolbox](https://www.mathworks.com/products/statistics.html) installed.
+We have recently added *Optimal Kmeans* algorithm to the `pop_clust` function. This feature allows you to find the optimal number of clusters for your data. To use this feature, you need to have the [MATLAB Statistics and Machine Learning Toolbox](https://www.mathworks.com/products/statistics.html) installed.
 
 To use this feature, you need to select the *Optimal Kmeans* option from the *Clustering algorithm* dropdown menu. Then, you need to input a range of cluster numbers to test (in the screenshot below minimum is set to 10 and maximum is set to 30). The algorithm will then test the clustering for each number of clusters in the range and chooses the optimal number of clusters based on the *silhouette* score. The *silhouette* score is a measure of how similar an object is to its own cluster compared to other clusters. The optimal number of clusters is the one that maximizes the *silhouette* score. Read more about the *silhouette* score from the [MATLAB documentation](https://www.mathworks.com/help/stats/clustering.evaluation.silhouetteevaluation.html).
 
-*Recommended number of clusters:* Following the estimated number of clusters above, we recommend setting the lower bound of the cluster range to the half the average number of components per subject. For example, if you have 20 components per subject, set the lower bound to 10. Similarly, set the upper bound to 1.5 times the average number of components per subject. For example, if you have 20 components per subject, set the upper bound to 30. If the returned number of clusters is at its lower or upper bound, consider expanding the range. We also strongly recommend using the option to separate outliers.
+*Recommended number of clusters:* Following the rationale for estimated number of clusters above, we recommend setting the lower bound of the cluster range to the half the average number of components per subject. For example, if there are 20 components per subject, set the lower bound to 10. Similarly, set the upper bound to 1.5 times the average number of components per subject. For example, for 20 components per subject, set the upper bound to 30. If the returned number of clusters is at its lower or upper bound, consider expanding the range. We also strongly recommend using the option to separate outliers.
 
-![](/assets/images/studyclust5.png)
+![](/assets/images/studyclust14.png)
 
 Other clustering methods
 -----------------