diff --git a/ccn2019.Rmd b/ccn2019.Rmd
index 7301fc5..51b0168 100644
--- a/ccn2019.Rmd
+++ b/ccn2019.Rmd
@@ -1,6 +1,5 @@
 ---
 title: "Unbiased N-Back"
-date: "5/12/2019"
 output:
   ioslides_presentation: default
   slidy_presentation: default
@@ -16,26 +15,30 @@
 load('./data/CL2015.RData')
 ```
 
-## Intro
-
-### Problem: 
+## Problem
  - local statistical properties of the n-back affect how we respond.
  - local vs. global properties, what matters the most?
 
-### By-products
+### 
  - an agent to replicate behavioral data sets
  - an online service to generate and evaluate n-back sequences
 
 ## Method
- - create history window (a.k.a, contiguous subsequences)
- - calculate local T, L, S, U, $RT_{mean}$, $Accuracy_{mean}$ for each subsequence
+ - create a history window for each trial (a.k.a, contiguous subsequences)
+ - calculate local $T$, $L$, $S$, $U$, $RT_{mean}$, $Accuracy_{mean}$ for each subsequence
  - Model RT/Acc (response vars) with local properties (exp. vars)
  - Cluster responses (or exp. vars?)
  - Investigate if extracted clusters are statistically different
 
-## Modeling
- - Create two models for local and global features as explanatory vars
- - Continue with modeling RT and Accuracy based upon local and global feats and compare them. Which model provides a better description of the recoreded RT and Accuracy vars? (model comparasion, model selection, etc)
+### Explanatory Variables
+- $T$ number of targets
+- $L$ number of lures
+- $S$ Skewness score
+- $U$ Uniformity (!repetition)
+
+### Response Variables
+- $RT_{mean}$
+- $Accuracy_{mean}$
 
 ## Constraints
 
@@ -44,6 +47,10 @@
 - uniform distribution of choices
 - controlled local lumpiness
 
+## Modeling
+ - Create two models for local and global features as explanatory vars
+ - Continue with modeling RT and Accuracy based upon local and global feats and compare them. Which model provides a better description of the recoreded RT and Accuracy vars? (model comparasion, model selection, etc)
+
 
 ```{r}
 trials <- c('a','b','c','d','c','d','b','a','a','d','b','a','c','c','a','c')
@@ -65,9 +72,11 @@
 ```
 
 
-Each constraint is a cost function to minimize for each sequence of stimuli
+Each constraint is a cost function to minimize for each sequence of stimuli.
 
-```
+```{r, eval=F}
+
+# Codes for fitness and loss functions
 history <- contig_seqs
 targets <- 4
 lures <- 2
@@ -106,47 +115,56 @@
 
 ```{r}
 
-#NB %>%
-#  filter(participant=='P1') %>%
-#  group_by(block) %>%
-#  summarise(trials=n_distinct(trial))
-
-with_lures <- function(stim, stim_type, history) {
+# Codes to calculate local statistical properties
+with_lures <- function(condition, stim, stim_type, history = NA) {
+  
   # extend to 2-back/3-back
-
-  res <- sapply(
+  sapply(
     1:length(stim), 
     function(i) {
-      ifelse(
-        stim[i]==stri_sub(history[i],-2,-2) ||  stim[i]==stri_sub(history[i],-4,-4), 
-        'lure', 
-        as.character(stim_type[i])
+      switch(condition[i],
+             "2-back" = {
+               ifelse(
+                 stim[i]==stri_sub(history[i],-2,-2) ||  stim[i]==stri_sub(history[i],-4,-4), 
+                 'lure', 
+                 as.character(stim_type[i])
+               )},
+             "3-back" = {
+               ifelse(
+                 stim[i]==stri_sub(history[i],-3,-3) ||  stim[i]==stri_sub(history[i],-5,-5), 
+                 'lure', 
+                 as.character(stim_type[i])
+               )}
       )
+
     })
-  as.factor(res)
 }
 
-with_targets_ratio <- function(correct, history = c(), block_size=NA) {
-  if (is.na(block_size)) block_size = str_length(history)
-  sapply(1:length(correct), function(i) {
-    0 #TODO
+with_targets_ratio <- function(stimulus_type, history) {
+  sapply(1:length(history), function(i) {
+    trials <- stimulus_type[(i-str_length(history[i])):i]
+    trials <- unlist(trials, use.names=FALSE)
+    length(trials[trials=="target"])
   })
 }
 
 with_lures_ratio <- function(stimulus_type, history) {
-  res <- sapply(1:length(history), function(i) {
+  sapply(1:length(history), function(i) {
     trials <- stimulus_type[(i-str_length(history[i])):i]
     trials <- unlist(trials, use.names=FALSE)
     length(trials[trials=="lure"])
   })
-  res
 }
 
-with_skewness_score <- function(history) {
-  sapply(1:length(history), function(i) 0)
+with_lumpiness_score <- function(stimulus, history) {
+  sapply(1:length(history), function(i) {
+    trials <- stimulus[(i-str_length(history[i])):i]
+    trials <- unlist(trials, use.names=FALSE)
+    max(table(trials)) - 1
+  })
 }
 
-with_lumpiness_score <- function(history) {
+with_skewness_score <- function(stimulus, history) {
   sapply(1:length(history), function(i) 0)
 }
 
@@ -159,6 +177,7 @@
 }
 
 normalize_scores <- function(targets_ratio, lures_ratio, skewness, lumpiness) {
+  #TODO
   sapply(1:length(targets_ratio), function(i) 0)
 }
 
@@ -166,25 +185,27 @@
   group_by(participant, condition, block) %>%
   mutate(history = with_history(stimulus)) %>%
   #mutate(stimulus_type = map_chr(.x=stimulus, stim_type=stimulus_type, history=history,.f=with_lures))
-  mutate(stimulus_type_2 = with_lures(stimulus, stimulus_type, history)) %>%
-  mutate(targets_ratio = with_targets_ratio(correct)) %>%
-  mutate(lures_ratio = with_lures_ratio(stimulus_type_2, history)) %>%
-  mutate(skewness = with_skewness_score(history)) %>%
-  mutate(lumpiness = with_lumpiness_score(history)) %>%
+  mutate(stimulus_type_2 = with_lures(condition, stimulus, stimulus_type, history)) %>%
+  mutate(t = with_targets_ratio(stimulus_type_2, history)) %>%
+  mutate(l = with_lures_ratio(stimulus_type_2, history)) %>%
+  mutate(s = with_skewness_score(stimulus, history)) %>%
+  mutate(u = with_lumpiness_score(stimulus, history)) %>%
   #normalize_scores(targets_ratio, lures_ratio, skewness, lumpiness) %>%
   ungroup()
 
 # print
-NB_modified %>%
-  filter(participant=='P1', lures_ratio>0) %>%
-  View()
+# NB_modified %>%
+#   filter(participant=='P1') %>%
+#   View()
+#   
   
 ```
 
-#TODO
-modified_NB <- NB %>%
-  mutate(constraint1=fitness1(history), constrain2=fitness2(history), constraint3=fitness(history))
-kmeans(NB)
-ggplot(kmeans$accuracy)
-ggplot(kmeans$rt)
+## TODO
+  - data %>% mutate(constraint1=fitness1(history), constrain2=fitness2(history),
+  - constraint3=fitness(history))
+  - kmeans(NB)
+  - ggplot(kmeans_clusters$accuracy) 
+  - ggplot(kmeans_clusters$rt)
+  -