Stress and Resilience: Implications for Depression and Anxiety

Jerrold F. Rosenbaum, MD; Jennifer M. Covino, MPA


December 29, 2005

In This Article

Memory and Fear

Anderson and colleagues[6] employed neuroimaging to examine biological systems involved in the suppression of unwanted memories. Our current understanding of memory formation and the suppression of unwanted memories has established the roles of the hippocampus and lateral prefrontal cortex. The hippocampus is essential for declarative memory formation and is activated when memory is formed. The lateral prefrontal cortex, on the other hand, has been established as essential for suppressing memory. One hypothesis proposes that the prefrontal cortex plays an important role in disengaging hippocampal activity when memories are to be suppressed. Anderson's study used a "think/no-think" paradigm and functional MRI (fMRI) to examine suppression of unwanted memories. The results led to a proposed neurobiological model of memory control, postulating an interactive process by which people can "suppress" memory. The process of repressing memories may facilitate adaptation and coping in response to a traumatic stressor.

Study participants were presented with paired words and, while undergoing fMRI, were asked to recall and think of the associated response (Respond Condition) or prevent the associated word from entering consciousness during presentation of the word (Suppression Condition). Results from this study showed greater activation during suppression in a network of brain regions that included: the bilateral, dorsolateral, and ventrolateral prefrontal cortex; anterior cingulate; the contiguous presupplementary motor area; lateral premotor area in the rostral portion of the dorsal premotor cortex; and the intraparietal sulcus. By contrast, reduced activation was noted in the bilateral frontal polar cortex; posterior insula; left parietal cortex; and bilateral cuneus. In addition, during the suppression condition, reduced activation of the hippocampus bilaterally was observed.

A separate study conducted at the MGH by Milad and colleagues[7] reported a significant finding involving the relative thickness of the ventromedial prefrontal cortex (vmPFC), specifically in the medial orbitofrontal cortex, correlating with extinction retention. In this study, investigators measured skin conductance response (SCR) and cortical thickness (via fMRI scans). This 2-day experiment examined 14 healthy volunteers (8 men and 6 women; aged 21-34 years) and used a fear conditioning and extinction procedure by which a neutral conditioned stimulus (CS) is paired with an aversive unconditioned stimulus (US). The conditioned stimulus used photographs of 2 different rooms; each room contained a lamp with a different colored lamp shade. The US was a 500-ms electric shock delivered through electrodes attached to the second and third fingers of the participant's hand.

By Day 2, participants with greater vmPFC thickness (specifically thicker right medial orbitofrontal cortex) demonstrated reduced SCRs to the CS. Just as previous studies have indicated deficiencies in the vmPFC in patients with posttraumatic stress disorder (PTSD), results of this study implied that variability in the thickness of the vmPFC across the population may play a role in risk and resilience factors associated with anxiety disorders.

A second study conducted by Milad and associates[8] used a similar 2-day fear conditioning and extinction protocol to further examine whether the context of the stimuli affected extinction recall in a group of healthy volunteers as measured by SCRs and fMRI. This study assessed 30 individuals (16 men and 14 women; aged 18-42 years) The paradigm consisted of photographs of 2 different rooms (CXs, CX+, and CX-, chosen randomly and counterbalanced between the 2 groups) each of which contained a lamp (CS) with either a blue or red lampshade for half of the group and either a blue or yellow lampshade for the other half (CS+, CS-; counterbalanced across the 2 groups). In addition, an unconditioned stimulus (US) consisted of a 500-ms electric shock delivered through electrodes attached to the second and third finger of the participant's dominant hand. Day 1 of the study consisted of 3 phases: the habituation phase (4 CS+ and 4 CS- were presented); the conditioning phase (5 CS+ and 5 CS-, all presented within a CX+ with the US phase occurring immediately after each CX+ offset; before this phase, participants were instructed that the experiment was about to begin, that they "may or may not be shocked" during this phase and the following phases, and that if they were to be shocked it would occur after the very end of a picture presentation block); and the extinction phase (comprising 2 subphases, early and late, which were separated by a 1-minute rest period; participants were again reminded that they "may or may not be shocked"; each subphase consisted of 5 CS+ and 5 CS- all presented with 1 CS- associated with US) On Day 2, participants were told that this part of the experiment would be the same as on Day 1 and to use their memory to predict the occurrence of a shock. In addition, 3 phases were added: a recall phase (identical to the extinction subphase given on Day 1); a renewal phase (5 CS+ and 5 CX- were presented within the CS+ with no US delivered); and lastly the reinstatement phase (2 18-s presentations of the CX+ without any CS with the US delivered at the CX+ offset followed by 5 CS+ and 5 CS- trials within the CS+, without any US, to test for the reinstatement of the conditioned response.

Participants were divided into 2 groups, an extinction group (n=20) and a no-extinction group (n=10); the no-extinction group underwent the same procedure except that the Day 1 extinction phase was omitted. Results from this study indicated that the skin conductance levels were significantly higher in the CS+ trials than in the CS- trials during trials 2 and 4 (P < .05). During the recall phase, for participants exposed to the CS+ and CS- in the extinction context (CX-), the change in skin conductance levels to the CS+ and CS- did not differ; however, during the renewal phase, when participants were exposed to the CS+ and CS- in the conditioning context (CX+), ANOVA revealed significant main effects for stimulus, F(1,19)=12.8, P < .01, and trials, F(4,76)=19.7, P < .01. In addition, during the reinstatement phase, no difference in responses to CS+ and CS- was observed; however, changes in skin conductance levels to both CS+ and CS- were slightly elevated during the first reinstatement trial. Post hoc comparisons revealed a significantly larger response to the CS+ than the CS- during the conditioning phase (P = .05), although this quickly diminished during the early and late extinction training phases (P = .05 for both phases). No difference was observed during the recall phase, yet post hoc analysis revealed that differential condition responses reappeared when participants were exposed to the CS+ in the CX+ (P < .05).

Post hoc comparison also showed that the mean response to the CS+ was significantly larger during the renewal phase (which excluded an US) than during the recall phase (which was associated with 1 US) (P < .05). Participants in the no-extinction group had relatively small changes in skin conductance response to the CS+ within the extinction context during the recall phase, suggesting that they were able to retain the extinction training that had taken place the previous day. This group experienced a significantly larger mean change in skin conductance during the recall phase to CS+ (t(28)=-3.9, P < .01), but the mean change in skin conductance level to CS+ during the recall phase (.037± 0.10) was not significantly different from its response during the conditioning phase (t(18)=1.6, P > .1). During the recall phase, the extinction group recovered only 32% of the conditioned response in the CX-, whereas the no-extinction group recovered 95% of the change in skin conductance when exposed to the CS+ in the CX+ during the renewal phase on Day 2 (t(19)=-5.1, P < .01). The percent of conditioned response recovered in the CX- during the recall phase by the no-extinction group was also significantly larger than that recovered by the extinction group (t(28)=3.8, P < .01).

The overall results of this study indicate that the expression of long-term memory for fear extinction is strongly influenced by contextual information, and exposure to such information may trigger a stress response. As indicated in this study, memory of a stressor (eg, the extinction of a conditioned fear response) was expressed upon presentation of the associated contextual information following a 24-hour delay period.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.