Raine et al (1997)

to use PET scans to identify brain impairments in murderers who had pleaded NGRI

41 murderers pleaded NGRI (39 men and 2 women) in California, USA + 41 controls.
matched pairs design for 3 confounding variables - age, sex + ethnicity (6 also matched for sz)

ppts free from medication in the 2 weeks running up to PET scans (confrimed by urine tests)
control ppts had a physical examination + psychiatric interview
PET scan - injected with radioactive tracer to light up metabolism on the scanner
each then completed a CPT for 32 mins (identifying targets on screen + pressing button)
PET scan conducted immediately after 32 minute period
PET scan took 10 images through the brain at 10mm intervals from the canthomeatal line

Murderers had lower glucose metabolism than controls in the left amygdala (feelings of happiness + emotional memory) and increased activity in the right amygdala (negative feelings such as fear)

reduced activity in PFC and aysmmetries in amygdala of NGRI murderers in comparison to controls

murderers pleading NGRI have different brain activity from people who are not violent offenders
the murderers had impaired functioning in areas of the brain previously associated with aggression
these impairments create a predisposition to violence that is only expressed in behaviour when social, environmental and psychological conditions are 'right'

low
unapplicable to murderers who pleaded not guilty or guilty
only applicable to this very specific group
experimental groujp had 39 men and 2 women - results could be androcentric and unreflective of women

one key element was not matched within Raine's study - 23 murderers had a history of head injury which may explain differences in activity levels between the two groups
decreased internal validity - may not have measured what he intended and therefore conclusions of the offenders having different brain activity may be incorrect and due to this confounding variable

high degree of control (ppts free from meds + lab experiment allowed for control over confounding variables)

decreased - results can't be applied to real life (PET scanning can cause stess + lab environment is unnatural)

standardised procedures (PET scans took 10 images in 10mm intervals and each ppt did the same CPT for 32 mins (identifying targets and pressing a button) with the PET scan following immediately after
allowing for easy replication to check for consistency in results of different brain activity in offenders causing aggression

PET scans take images in 10mm slices relative to the canthomeatal line (the line from outer corner of eye to middle ear) but this line is imaginary and is known to vary significantly between individuals. this means it is difficult to locate specific brain areas. This decreases internal validity as the measurements may have been from entirely different brain structures than what was reported, leading to inaccurate findings about differences in brain activity within the violent offenders

Raine does not suggest that since brain structures that can cause aggression can be identified that it can be ‘treated’ by medication or surgery, instead arguing that such treatments wouldn’t be justified because brain structures is only a partial explanation.
- However the study provides useful preliminary evidence and paves the way for future research.

- aim
- sample / generalisabilty / brain injury not matched
- procedure / internal validity / external validity
- procedure / reliability / canthomeatal line
- conclusion / application