Pavlovian fear conditioning is a behavioral paradigm in which organisms learn to predict aversive events.[1] It is a form of learning in which an aversive stimulus (e.g. an electrical shock) is associated with a particular neutral context (e.g., a room) or neutral stimulus (e.g., a tone), resulting in the expression of fear responses to the originally neutral stimulus or context. This can be done by pairing the neutral stimulus with an aversive stimulus (e.g., an electric shock, loud noise, or unpleasant odor[2]). Eventually, the neutral stimulus alone can elicit the state of fear. In the vocabulary of classical conditioning, the neutral stimulus or context is the "conditional stimulus" (CS), the aversive stimulus is the "unconditional stimulus" (US), and the fear is the "conditional response" (CR).
Fear conditioning has been studied in numerous species, from snails[3] to humans.[4] In humans, conditioned fear is often measured with verbal report and galvanic skin response. In other animals, conditioned fear is often measured with freezing (a period of watchful immobility) or fear potentiated startle (the augmentation of the startle reflex by a fearful stimulus). Changes in heart rate, breathing, and muscle responses via electromyography can also be used to measure conditioned fear. A number of theorists have argued that conditioned fear coincides substantially with the mechanisms, both functional and neural, of clinical anxiety disorders.[5] Research into the acquisition, consolidation and extinction of conditioned fear promises to inform new drug based and psychotherapeutic treatments for an array of pathological conditions such as dissociation, phobias and post-traumatic stress disorder.[6]
Scientists have discovered that there is a set of brain connections that determine how fear memories are stored and recalled. While studying rats' ability to recall fear memories, researchers found a newly identified brain circuit is involved. Initially, the pre-limbic prefrontal cortex (PL) and the basolateral amygdala (BLA) were identified in memory recall. A week later, the central amygdala (CeA) and the paraventricular nucleus of the thalamus (PVT) were identified in memory recall, which are responsible for maintaining fear memories. This study shows how there are shifting circuits between short term recall and long term recall of fear memories. There is no change in behavior or response, only change in where the memory was recalled from.[7]
In addition to the amygdala, the hippocampus and the anterior cingulate cortex are important in fear conditioning. Fear conditioning in the rat is stored at early times in the hippocampus, with alterations in hippocampal gene expression observed at 1 hour and 24 hours after the event.[8] In the mouse, changed gene expression is also seen in the hippocampus at one hour and 24 hours after fear conditioning. These changes are transient in the hippocampal neurons, and almost none are present in the hippocampus after four weeks. By 4 weeks after the event, the memory of the fear conditioning event is more permanently stored in the anterior cingulate cortex.[9][10][11]
Neurobiology
editNeuronal gene expression
editAs shown in the rodent brain, neuronal gene expression is dynamically changed in response to fear conditioning. In particular, the expressions of immediate early genes (IEGs) such as Egr1, c-Fos, and Arc are rapidly and selectively up-regulated in subsets of neurons in specific brain regions associated with learning and memory formation.[12]
A review in 2022[13] describes multiple steps in up-regulating the IEGs in neurons in the hippocampus during fear conditioning. IEGs are similarly up-regulated in the amygdala during fear conditioning.[14][15] The multiple steps in up-regulating IEGs[13] include activation of transcription factors,[16] formation of chromatin loops, interaction of enhancers with promoters in chromatin loops and topoisomerase II beta-initiated temporary DNA double-strand breaks.
At least two IEGs up-regulated by fear conditioning, Egr1 and Dnmt3A2 (shown to be an IEG by Oliveira et al.[17]) affect DNA methylation, and thus expression, of many genes. Up-regulated EGR1 proteins associate with pre-existing nuclear TET1 proteins, and the EGR1 proteins bring TET1 proteins to hundreds of genes, allowing TET1 to initiate DNA demethylation of those genes.[18] DNMT3A2 protein is a de novo DNA methyltransferase, adding methylation to cytosines in DNA. Expression of DNMT3A2 proteins in hippocampus neurons in culture preferentially _targeted the addition of new methylation to more than 200 genes involved in synaptic plasticity.[19] Expressions of IEGs are a source of the dynamic changes in subsequent neuronal gene expression in response to fear conditioning.
Amygdala
editFear conditioning is thought to depend upon an area of the brain called the amygdala. The amygdala is involved in acquisition, storage, and expression of conditioned fear memory.[11] Lesion studies have revealed that lesions drilled into the amygdala before fear conditioning prevent the acquisition of the conditioned response of fear, and lesions drilled in the amygdala after conditioning cause conditioned responses to be forgotten.[20] Electrophysiological recordings from the amygdala have demonstrated that cells in that region undergo long-term potentiation (LTP), a form of synaptic plasticity believed to underlie learning.[21] Pharmacological studies, synaptic studies, and human studies also implicate the amygdala as chiefly responsible for fear learning and memory.[11] Additionally, inhibition of neurons in the amygdala disrupts fear acquisition, while stimulation of those neurons can drive fear-related behaviors, such as freezing behavior in rodents.[22] This indicates that proper function of the amygdala is both necessary for fear conditioning and sufficient to drive fear behaviors. The amygdala is not exclusively the fear center, but also an area for responding to various environmental stimuli. Several studies have shown that when faced with unpredictable neutral stimuli, amygdala activity increases. Therefore, even in situations of uncertainty and not necessarily fear, the amygdala plays a role in alerting other brain regions that encourage safety and survival responses.[23]
In the mid-1950s Lawrence Weiskrantz demonstrated that monkeys with lesions of amygdala failed to avoid an aversive shock while the normal monkeys learned to avoid them. He concluded that a key function of the amygdala was to connect external stimuli with aversive consequences.[24] Following Weiskrantz's discovery many researchers used avoidance conditioning to study neural mechanisms of fear.[25]
Joseph E. LeDoux has been instrumental in elucidating the amygdala's role in fear conditioning. He was one of the first to show that the amygdala undergoes long-term potentiation during fear conditioning, and that ablation of amygdala cells disrupts both learning and expression of fear.[26]
Hippocampus
editSome types of fear conditioning (e.g. contextual and trace) also involve the hippocampus, an area of the brain believed to receive affective impulses from the amygdala and to integrate those impulses with previously existing information to make it meaningful. Some theoretical accounts of traumatic experiences suggest that amygdala-based fear bypasses the hippocampus during intense stress and can be stored somatically or as images that can return as physical symptoms or flashbacks without cognitive meaning.[27]
The hippocampus is one of the brain regions that undergoes major alterations in gene expression after contextual fear conditioning. Contextual fear conditioning applied to a rat causes about 500 genes to be up-regulated (possibly due to DNA demethylation of CpG sites) and about 1,000 genes to be down-regulated (observed to be correlated with DNA methylation at CpG sites in promoter regions) (see Regulation of transcription in learning and memory).[8] By 24 hours after the event, 9.17% of the genes in the genomes of rat hippocampus neurons are differentially methylated. The pattern of induced and repressed genes within hippocampal neurons appears to provide a molecular basis for forming the early transient memory of contextual fear conditioning in the hippocampus.[8] When similar contextual fear conditioning was applied to a mouse, one hour after contextual fear conditioning there were 675 demethylated genes and 613 hypermethylated genes in the hippocampus region of the mouse brain.[9] These changes were transient in the hippocampal neurons, and almost none of these DNA methylation alterations were present in the hippocampus after four weeks. However, in mice subjected to contextual fear conditioning, after four weeks there were more than 1,000 differentially methylated genes and more than 1,000 differentially expressed genes in the mouse anterior cingulate cortex[9] where long-term memories are stored.
Molecular mechanisms
editDouble-strand breaks
editMore than 100 DNA double-strand breaks occur, both in the hippocampus and in the medial prefrontal cortex (mPFC), in two peaks, at 10 minutes and at 30 minutes after contextual fear conditioning.[28] This appears to be earlier than the DNA methylations and demethylations of neuron DNA in the hippocampus that were measured at one hour and 24 hours after contextual fear conditioning (described above in the section Hippocampus).
The double strand breaks occur at known memory-related immediate early genes (among other genes) in neurons after neuron activation.[29][28] These double-strand breaks allow the genes to be transcribed and then translated into active proteins.
One immediate early gene newly transcribed after a double-strand break is EGR1. EGR1 is an important transcription factor in memory formation. It has an essential role in brain neuron epigenetic reprogramming. EGR1 recruits the TET1 protein that initiates a pathway of DNA demethylation. Removing DNA methylation marks allows the activation of downstream genes (see Regulation of gene expression#Regulation of transcription in learning and memory. EGR1 brings TET1 to promoter sites of genes that need to be demethylated and activated (transcribed) during memory formation.[30] EGR-1, together with TET1, is employed in programming the distribution of DNA demethylation sites on brain DNA during memory formation and in long-term neuronal plasticity.[30]
DNMT3A2 is another immediate early gene whose expression in neurons can be induced by sustained synaptic activity.[17] DNMTs bind to DNA and methylate cytosines at particular locations in the genome. If this methylation is prevented by DNMT inhibitors, then memories do not form.[31] If DNMT3A2 is over-expressed in the hippocampus of young adult mice it converts a weak learning experience into long-term memory and also enhances fear memory formation.[32]
Intra-amygdala circuit
editNeurons in the basolateral amygdala are responsible for the formation of conditioned fear memory. These neurons project to neurons in the central amygdala for the expression of a conditioned fear response. Damage to these areas in the amygdala would result in disruption of the expression of conditioned fear responses. Lesions in the basolateral amygdala have shown severe deficits in the expression of conditioned fear responses. Lesions in the central amygdala have shown mild deficits in the expression of conditioned fear responses.[11]
NMDA receptors and glutamate
editOne of the major neurotransmitters involved in conditioned fear learning is glutamate.[33] It has been suggested that NMDA receptors (NMDARs) in the amygdala are necessary for fear memory acquisition, because disruption of NMDAR function disrupts development of fear responses in rodents.[33] In addition, the associative nature of fear conditioning is reflected in the role of NMDARs as coincident detectors, where NMDAR activation requires simultaneous depolarization by US inputs combined with concurrent CS activation.[34]
Epigenetics
editConditioned fear may be inherited transgenerationally. In one experiment, mice were conditioned to fear an acetophenone odor and then set up to breed subsequent generations of mice. Those subsequent generations of mice also showed a behavioral sensitivity to acetophenone, which was accompanied by neuroanatomical and epigenetic changes that are believed to have been inherited from the parents' gametes.[35]
Across development
editThe learning involved in conditioned fear, as well as the underlying neurobiology, changes dramatically from infancy, across childhood and adolescence, into adulthood and aging. Specifically, infant animals show an inability to develop fear associations, whereas their adult counterparts develop fear memories much more readily.[36]
Previous research has indicated that adolescents show hampered fear extinction learning compared to children and adults.[37] This finding may have clinical implications, as one of the most widely used treatments for anxiety disorders is exposure based therapy, which builds on the principles of fear extinction. The exact mechanisms underlying the developmental differences in fear extinction learning have not yet been discovered, although it has been suggested that age related differences in connectivity between the amygdala and medial prefrontal cortex can be one of the biological mechanisms underpinning the developmental change in fear extinction learning.[38]
Prior experience with stress
editA history of stressors preceding a traumatic event increases the effect of fear conditioning in rodents.[39] This phenomenon, named Stress-Enhanced Fear Learning (SEFL), has been demonstrated in both young (e.g. Poulos et al. 2014[40]) and adult (e.g. Rau et al. 2009[41]) rodents. Biological mechanisms underpinning SEFL have not yet been made clear, though it has been associated with a rise in corticosterone, the stress hormone, following the initial stressor.[42]
See also
editReferences
edit- ^ Maren S (2001). "Neurobiology of Pavlovian fear conditioning". Annual Review of Neuroscience. 24: 897–931. doi:10.1146/annurev.neuro.24.1.897. hdl:2027.42/61939. PMID 11520922.
- ^ Wallace KJ, Rosen JB (October 2000). "Predator odor as an unconditioned fear stimulus in rats: elicitation of freezing by trimethylthiazoline, a component of fox feces". Behavioral Neuroscience. 114 (5): 912–22. doi:10.1037/0735-7044.114.5.912. PMID 11085605.
- ^ Walters ET, Carew TJ, Kandel ER (January 1981). "Associative Learning in Aplysia: evidence for conditioned fear in an invertebrate". Science. 211 (4481): 504–6. Bibcode:1981Sci...211..504W. doi:10.1126/science.7192881. PMID 7192881.
- ^ Critchley HD, Mathias CJ, Dolan RJ (February 2002). "Fear conditioning in humans: the influence of awareness and autonomic arousal on functional neuroanatomy". Neuron. 33 (4): 653–63. doi:10.1016/s0896-6273(02)00588-3. PMID 11856537.
- ^ Rosen JB, Schulkin J (April 1998). "From normal fear to pathological anxiety". Psychological Review. 105 (2): 325–50. doi:10.1037/0033-295X.105.2.325. PMID 9577241.
- ^ VanElzakker MB, Dahlgren MK, Davis FC, Dubois S, Shin LM (September 2014). "From Pavlov to PTSD: the extinction of conditioned fear in rodents, humans, and anxiety disorders". Neurobiology of Learning and Memory. 113: 3–18. doi:10.1016/j.nlm.2013.11.014. PMC 4156287. PMID 24321650.
- ^ Yeager A (19 January 2015). "Newly identified brain circuit hints at how fear memories are made" (PDF). Science News.
- ^ a b c Duke CG, Kennedy AJ, Gavin CF, Day JJ, Sweatt JD (July 2017). "Experience-dependent epigenomic reorganization in the hippocampus". Learn Mem. 24 (7): 278–288. doi:10.1101/lm.045112.117. PMC 5473107. PMID 28620075.
- ^ a b c Halder R, Hennion M, Vidal RO, Shomroni O, Rahman RU, Rajput A, Centeno TP, van Bebber F, Capece V, Garcia Vizcaino JC, Schuetz AL, Burkhardt S, Benito E, Navarro Sala M, Javan SB, Haass C, Schmid B, Fischer A, Bonn S (January 2016). "DNA methylation changes in plasticity genes accompany the formation and maintenance of memory". Nat Neurosci. 19 (1): 102–10. doi:10.1038/nn.4194. PMID 26656643. S2CID 1173959.
- ^ Frankland PW, Bontempi B, Talton LE, Kaczmarek L, Silva AJ (May 2004). "The involvement of the anterior cingulate cortex in remote contextual fear memory". Science. 304 (5672): 881–3. Bibcode:2004Sci...304..881F. doi:10.1126/science.1094804. PMID 15131309. S2CID 15893863.
- ^ a b c d Kim JJ, Jung MW (2006). "Neural circuits and mechanisms involved in Pavlovian fear conditioning: a critical review". Neurosci Biobehav Rev. 30 (2): 188–202. doi:10.1016/j.neubiorev.2005.06.005. PMC 4342048. PMID 16120461.
- ^ Minatohara K, Akiyoshi M, Okuno H (2015). "Role of Immediate-Early Genes in Synaptic Plasticity and Neuronal Ensembles Underlying the Memory Trace". Front Mol Neurosci. 8: 78. doi:10.3389/fnmol.2015.00078. PMC 4700275. PMID 26778955.
- ^ a b Bernstein C (2022). "DNA Methylation and Establishing Memory". Epigenet Insights. 15: 25168657211072499. doi:10.1177/25168657211072499. PMC 8793415. PMID 35098021.
- ^ Rosen JB, Fanselow MS, Young SL, Sitcoske M, Maren S (June 1998). "Immediate-early gene expression in the amygdala following footshock stress and contextual fear conditioning". Brain Res. 796 (1–2): 132–42. doi:10.1016/s0006-8993(98)00294-7. hdl:2027.42/56231. PMID 9689463.
- ^ Martinez RC, Gupta N, Lázaro-Muñoz G, Sears RM, Kim S, Moscarello JM, LeDoux JE, Cain CK (July 2013). "Active vs. reactive threat responding is associated with differential c-Fos expression in specific regions of amygdala and prefrontal cortex". Learn Mem. 20 (8): 446–52. doi:10.1101/lm.031047.113. PMC 3718200. PMID 23869027.
- ^ Bahrami S, Drabløs F (September 2016). "Gene regulation in the immediate-early response process". Adv Biol Regul. 62: 37–49. doi:10.1016/j.jbior.2016.05.001. PMID 27220739.
- ^ a b Oliveira AM, Hemstedt TJ, Bading H (July 2012). "Rescue of aging-associated decline in Dnmt3a2 expression restores cognitive abilities". Nat Neurosci. 15 (8): 1111–3. doi:10.1038/nn.3151. PMID 22751036. S2CID 10590208.
- ^ Sun Z, Xu X, He J, Murray A, Sun MA, Wei X, Wang X, McCoig E, Xie E, Jiang X, Li L, Zhu J, Chen J, Morozov A, Pickrell AM, Theus MH, Xie H. EGR1 recruits TET1 to shape the brain methylome during development and upon neuronal activity. Nat Commun. 2019 Aug 29;10(1):3892. doi: 10.1038/s41467-019-11905-3. PMID 31467272
- ^ Gulmez Karaca K, Kupke J, Brito D, Zeuch B, Thome C, Weichenhan D, Lutsik P, Plass C, Oliveira A (January 2020). "Neuronal ensemble-specific DNA methylation strengthens engram stability". Nat Commun. 11 (1): 639. Bibcode:2020NatCo..11..639G. doi:10.1038/s41467-020-14498-4. PMC 6994722. PMID 32005851.
- ^ Maren S (1998). "Neurotoxic Basolateral Amygdala Lesions Impair Learning and Memory But Not the Performance of Conditional Fear in Rats". The Journal of Neuroscience. 19 (19): 8696–9703. doi:10.1523/JNEUROSCI. hdl:1842/36564. PMC 6783031. PMID 10493770.
- ^ Sah P, Westbrook RF, Lüthi A (2008-05-01). "Fear conditioning and long-term potentiation in the amygdala: what really is the connection?". Annals of the New York Academy of Sciences. 1129 (1): 88–95. Bibcode:2008NYASA1129...88S. doi:10.1196/annals.1417.020. PMID 18591471. S2CID 36506768.
- ^ Bocchio M, Nabavi S, Capogna M (May 2017). "Synaptic Plasticity, Engrams, and Network Oscillations in Amygdala Circuits for Storage and Retrieval of Emotional Memories". Neuron. 94 (4): 731–743. doi:10.1016/j.neuron.2017.03.022. PMID 28521127.
- ^ Grupe, D. W., & Nitschke, J. B. (2011). Anxiety disorders and the amygdala. In AccessScience. McGraw-Hill Education. doi:10.1036/1097-8542.YB110087
- ^ Weiskrantz L (August 1956). "Behavioral changes associated with ablation of the amygdaloid complex in monkeys". Journal of Comparative and Physiological Psychology. 49 (4): 381–91. doi:10.1037/h0088009. PMID 13345917.
- ^ Kandel ER, Schwartz JH, Jessel TH, Siegelbaum SA, Hudspeth AJ (2013). Principles of neural science. United States of America: McGraw Hill Medical. p. 1084. ISBN 978-0-07-139011-8.
- ^ LeDoux JE (2000-03-01). "Emotion circuits in the brain". Annual Review of Neuroscience. 23 (1): 155–84. doi:10.1146/annurev.neuro.23.1.155. PMID 10845062.
- ^ Bromberg PM (2003). "Something wicked this way comes: Trauma, dissociation, and conflict: The space where psychoanalysis, cognitive science, and neuroscience overlap". Psychoanalytic Psychology. 20 (3): 558–74. doi:10.1037/0736-9735.20.3.558.
- ^ a b Stott RT, Kritsky O, Tsai LH (2021). "Profiling DNA break sites and transcriptional changes in response to contextual fear learning". PLOS ONE. 16 (7): e0249691. Bibcode:2021PLoSO..1649691S. doi:10.1371/journal.pone.0249691. PMC 8248687. PMID 34197463.
- ^ Madabhushi R, Gao F, Pfenning AR, Pan L, Yamakawa S, Seo J, Rueda R, Phan TX, Yamakawa H, Pao PC, Stott RT, Gjoneska E, Nott A, Cho S, Kellis M, Tsai LH (June 2015). "Activity-Induced DNA Breaks Govern the Expression of Neuronal Early-Response Genes". Cell. 161 (7): 1592–605. doi:10.1016/j.cell.2015.05.032. PMC 4886855. PMID 26052046.
- ^ a b Sun Z, Xu X, He J, Murray A, Sun MA, Wei X, Wang X, McCoig E, Xie E, Jiang X, Li L, Zhu J, Chen J, Morozov A, Pickrell AM, Theus MH, Xie H (August 2019). "EGR1 recruits TET1 to shape the brain methylome during development and upon neuronal activity". Nat Commun. 10 (1): 3892. Bibcode:2019NatCo..10.3892S. doi:10.1038/s41467-019-11905-3. PMC 6715719. PMID 31467272.
- ^ Bayraktar G, Kreutz MR (April 2018). "Neuronal DNA Methyltransferases: Epigenetic Mediators between Synaptic Activity and Gene Expression?". Neuroscientist. 24 (2): 171–185. doi:10.1177/1073858417707457. PMC 5846851. PMID 28513272.
- ^ Oliveira AM (October 2016). "DNA methylation: a permissive mark in memory formation and maintenance". Learn Mem. 23 (10): 587–93. doi:10.1101/lm.042739.116. PMC 5026210. PMID 27634149.
- ^ a b Johansen JP, Cain CK, Ostroff LE, LeDoux JE (October 2011). "Molecular mechanisms of fear learning and memory". Cell. 147 (3): 509–24. doi:10.1016/j.cell.2011.10.009. PMC 3215943. PMID 22036561.
- ^ Johansen JP, Hamanaka H, Monfils MH, Behnia R, Deisseroth K, Blair HT, LeDoux JE (July 2010). "Optical activation of lateral amygdala pyramidal cells instructs associative fear learning". Proceedings of the National Academy of Sciences of the United States of America. 107 (28): 12692–7. Bibcode:2010PNAS..10712692J. doi:10.1073/pnas.1002418107. PMC 2906568. PMID 20615999.
- ^ Dias BG, Ressler KJ (January 2014). "Parental olfactory experience influences behavior and neural structure in subsequent generations". Nature Neuroscience. 17 (1): 89–96. doi:10.1038/nn.3594. PMC 3923835. PMID 24292232.
- ^ Ganella DE, Kim JH (October 2014). "Developmental rodent models of fear and anxiety: from neurobiology to pharmacology". British Journal of Pharmacology. 171 (20): 4556–74. doi:10.1111/bph.12643. PMC 4209932. PMID 24527726.
- ^ Pattwell, Siobhan S.; Duhoux, Stéphanie; Hartley, Catherine A.; Johnson, David C.; Jing, Deqiang; Elliott, Mark D.; Ruberry, Erika J.; Powers, Alisa; Mehta, Natasha; Yang, Rui R.; Soliman, Fatima (2012-10-02). "Altered fear learning across development in both mouse and human". Proceedings of the National Academy of Sciences. 109 (40): 16318–16323. Bibcode:2012PNAS..10916318P. doi:10.1073/pnas.1206834109. ISSN 0027-8424. PMC 3479553. PMID 22988092.
- ^ Gee, Dylan G.; Humphreys, Kathryn L.; Flannery, Jessica; Goff, Bonnie; Telzer, Eva H.; Shapiro, Mor; Hare, Todd A.; Bookheimer, Susan Y.; Tottenham, Nim (2013-03-06). "A Developmental Shift from Positive to Negative Connectivity in Human Amygdala–Prefrontal Circuitry". Journal of Neuroscience. 33 (10): 4584–4593. doi:10.1523/JNEUROSCI.3446-12.2013. ISSN 0270-6474. PMC 3670947. PMID 23467374.
- ^ Sneddon EA, Riddle CA, Schuh KM, Quinn JJ, Radke AK (January 2021). "Selective enhancement of fear learning and resistance to extinction in a mouse model of acute early life trauma". Learning & Memory. 28 (1): 12–16. doi:10.1101/lm.052373.120. PMC 7747648. PMID 33323497.
- ^ Poulos AM, Reger M, Mehta N, Zhuravka I, Sterlace SS, Gannam C, et al. (August 2014). "Amnesia for early life stress does not preclude the adult development of posttraumatic stress disorder symptoms in rats". Biological Psychiatry. 76 (4): 306–14. doi:10.1016/j.biopsych.2013.10.007. PMC 3984614. PMID 24231200.
- ^ Rau V, Fanselow MS (March 2009). "Exposure to a stressor produces a long lasting enhancement of fear learning in rats". Stress. 12 (2): 125–33. doi:10.1080/10253890802137320. PMID 18609302. S2CID 15453890.
- ^ Beylin AV, Shors TJ (January 2003). "Glucocorticoids are necessary for enhancing the acquisition of associative memories after acute stressful experience". Hormones and Behavior. 43 (1): 124–131. doi:10.1016/S0018-506X(02)00025-9. PMC 3363955. PMID 12614642.