I found a paper https://www.ncbi.nlm.nih.gov/pubmed/21459521  
Med Hypotheses. 2011 Jun;76(6):769-73. doi: 10.1016/j.mehy.2011.01.020. Epub 2011 Apr 2.Tardive dysphoria: the role of long term antidepressant use in-inducing chronic depression.El-Mallakh RS1, Gao Y, Jeannie Roberts R.

This provoked me to inquire of my colleagues as to why there'd been no further exploration of this issue to be found in teh literature since 2011 and I got this reply from Robert Thatcher a neuroscientist and developer of the Neuroguide EEG analysis software that we use at Quietmind for our clnical and research activities. It is clear that the continued use of SSRI medications is now getting more critical attention now 30 years after the arrival of Prozac. Neurofeedback and other treatment methods are able to provide similar results without side effects and that last long after the training process has been completed. Here's Dr. Thatcher's reply:

This post concerns he science and publications showing negative effects after long term use of antidepressants.  The negative effect of the antidepresssants is called Tardive Dysphoria which is chronic depression induced by long term use of antidepressants.  There are several interesting studies on this topic if one does a simple Google search as well as a search of the National Library of Medicine database, e.g., https://www.ncbi.nlm.nih.gov/pubmed/30199999Tardive dysphoria: the role of long term antidepressant use in-inducing chronic depression
http://unthinkable.cc/long-term-evidence-we-cant-ignore-anymore-anti-depressant-outcomes/
http://behaviorismandmentalhealth.com/2014/04/08/antidepressants-make-things-worse-in-the-long-term/

A common hypothesis is that anti-depressants inhibit re-uptake that results in increased serotonin in the extracellular space however over time there are compensatory mechanisms that remove serotonin faster than the drug can block re-uptake or different re-uptake mechanisms develop that result in a worse and chronic outcome.

Ketamine has a different mode of action, e.g., midline thalamus de-coupling to the medial frontal lobes and hippocampus that appears to disrupt a cyclic dynamic and sort of refreshes the circuity for a while,e.g., Cortical-brainstem (nu. Raphe) regulatory control mechanisms.   A search of the National Library of Medicine database using the search terms "Depression Brain Network" results in 2,875 citations.  There is good agreement about the hubs of this network between different imaging modalities such as fMRI, PET and EEG/MEG.  Therefore, a reasonable idea is to measure the brain networks related to depression using QEEG and LORETA as a baseline assessment and then use NFB to reinforce increased stability and efficiency of the nodes and connections between the hubs of the depression network as well as to use QEEG to assess the relative effectiveness of non-SSRI treatment..   As Wes stated, exercise and good diet (Typtophan is a precursor of serotonin) and lifestyle changes and perhaps also Ketamine.  However, given the risk of doing harm based on the over use of anti-depressant drugs as published in the scientific literature points toward considering alternative treatments.  
Bob Thatcher, PhD  Applied Neuroscience Inc.