Dissertation Results Comparisons - which SPSS method should I use?

#1
Hi everyone,

I'm in my final year as a Psychology student aspiring to be a cognitive neuroscientist (accepted to birmingham university provided I get my 2.1 class honours :D), firstly I'd like to thank you for your time to read this (especially since there is a lot of reading to be done, I apologize), secondly I'll be describing my study and then moving on to how I planned to present my results. All the tables and graphs I talk about will be attached :)

All hypotheses etc are on the attached document :).

So, my dissertation topic is "A comparison of EEG oscillations between a sample of children with ADHD and a non-ADHD control group during a visual memory task." More specifically, the aim of this study is to determine the effects of methylphenidate (Ritalin) on concentration.

The experiment consists of analysis of the Parietal lobes (during EEG experimentation). I will be focusing on the Left and Right parietal lobe separately.

The children in the experiment are subject to 2 types of tasks.

1) Sensory Co-ordination Task
2) Sensory Distraction Task.

The children are divided into 3 groups for each of the tasks.

1) ADHD medicated group (ingestion of ritalin tablet 2 hours prior to experiment)
2) ADHD non-medicated group (same children, just after the effects have worn off)
3) Control group of healthy children.

After each group carries out both tasks the EEG results are presented as Frequency / Time graphs with a dB colour scale on the right hand side, which is used to discern the decibels of the brain waves. (blue to red, red being highest). (I believe details of the task are irrelevant since due to ethical issues, I cannot have the individual results of each child. Instead, I have the results as averaged graphs.)

The graphs themselves are coloured, to analyze this and compare them quantitatively, I thought it best to turn them into a table method of display.

I did this by dividing each graph into 24 sections. This was done by drawing vertical lines from 100 / 200 / 300 / 400 ms on the x-axis and horizontal lines from the y-axis
on 5 / 10 / 15/ 20 / 25 / 30 hz. On each intersection point of these lines, I would compare the colour to the dB colour scale and write down the value. This value was then inputted into the table (Columns = ms, Rows = hz). After the tables are filled in with the values of dB intensity (colour) each value is summed up to give 1 total value for each table. I believe this would make it easier to compare them to each other.

Seeing as the more red the colour is, the higher the dB it stands to reason that the higher the sum of each table the more brain activity therefore the more concentration, right? (hope so :( )

So basicaly that is how I have presented my results thus far. Now with regards to comparisons, I have to compare ADHD MED group to control group, and ADHD NON MED group to control group. This has to be done for each task for each lobe.

Now it is important to say that I DO NOT HAVE THE INDIVIDUAL SCORES OF EACH CHILD due to ethic concerns....I JUST HAVE THE AVERAGES FROM THE GRAPH. I have tried ANOVA, T-TESTS, CHI-SQUARED (which I have learned in my statistics study unit) but I simply do not seem to have the required data in order to input into each test :(.

So, I will summarize the process again just to make sure that it is understood as this is quite important:

-3 groups, adhd med, adhd non med, control. Each carrying out 2 tasks.
-Parietal Lobes (Left/Right)
-Sensory Distraction task/sensory Co-ordination task for each lobe.
-Graphs.
-Tables derived from graphs.
-all values in the table summed up to give 1 number, which is to be compared with that of the other groups.
-ADHD med compared to control, and ADHD non-med compared to control for each task for each parietal lobe.

I have explained this as best I can, english is my second language so I hope that this is good enough. Ideally everything is left as is and some kind soul can provide me with a statistical comparison method to compare my results as described above. I understand that this may be impossible so if you have any suggestions regarding any possible tweaks for the way I can present my results I am more than happy to listen to what you have to say.

The end, you made it, congratulations! :) I'd like to wish you all a Happy Easter, and thanks so much for your time :) I hope someone could help as my entire future depends on this experiment!
 
#2
1. Welcome aboard :)
2. I envy you! Congrats on neuroscience. :)
3. A mixed model should be conducted. The random factors are the repeated patients and lobes.
 
#3
Thanks for the quick reply :)

I haven't done mixed models in my statistics study unit, so I'm trying to learn how on youtube and I'm finding many versions such as mixed models anova, mixed models t-test? I hope I'm not asking too much of you but, which one specifically would be best?

Also, "random factors are the repeated patients and lobes" I'm not sure what you mean by that seeing as the parietal lobes do not influence the table in any way...would just like to clarify that I do not have the patients individual results, all the results I have are just in the form of the graphs originally.

Thanks again and sorry for the confusion.