Filling gaps in data points when interpolating with missing data

I have sensor data that is logged in every 15 minutes, (96 data points per day, interpolated data). I have a lot of missing data points, and I want (when the missing data points is larger then 2 hours) to find the first missing data point and the last missing data point and replace every single data point in between with NaN.

I have tried to find the gaps in the data by using find(diff>2*60*1000) and to replace it with NaN but this will only account for one data point. I would be graetfull if somebody could help. I have attached sample data.

[KSSV]

Read about fill missing.