Differential effects of early hippocampal pathology on episodic and semantic memory

Science

Volumn 277, Issue 5324, 1997, Pages 376-380

  Vargha Khadem, F ^a   Gadian, D G ^a   Watkins, K E ^a   Connelly, A ^a   Van Paesschen, W ^a,b   Mishkin, M ^c  

^a UNIVERSITY COLLEGE LONDON   (United Kingdom)

^b UNIVERSITY COLLEGE LONDON   (United Kingdom)

^c NATIONAL INSTITUTE OF MENTAL HEALTH   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

ANTEROGRADE AMNESIA; ARTICLE; COGNITION; HIPPOCAMPUS; HUMAN; LANGUAGE; MEMORY; NUCLEAR MAGNETIC RESONANCE; PRIORITY JOURNAL; SPEECH;

ADOLESCENT; ADULT; AMNESIA; BRAIN MAPPING; CEREBRAL CORTEX; CHILD; FEMALE; HIPPOCAMPUS; HUMANS; MAGNETIC RESONANCE IMAGING; MALE; MEMORY; NEUROPSYCHOLOGICAL TESTS; TEMPORAL LOBE;

EID: 0030803727     PISSN: 00368075     EISSN: None     Source Type: Journal    
DOI: 10.1126/science.277.5324.376     Document Type: Article

References (42)

1. L. R. Squire and B. J. Knowlton, in The Cognitive Neurosciences, M. S. Gazzaniga, Ed. (MIT Press, Cambridge, MA, 1995), pp. 825-837; S. Zola-Morgan, L. R. Squire, S. J. Ramus, Hippocampus 4, 483 (1994).
2. L. R. Squire and B. J. Knowlton, in The Cognitive Neurosciences, M. S. Gazzaniga, Ed. (MIT Press, Cambridge, MA, 1995), pp. 825-837; S. Zola-Morgan, L. R. Squire, S. J. Ramus, Hippocampus 4, 483 (1994).
3. F. Wood, V. Ebert, M. Kinsboume, in Human Memory and Amnesia, L. S. Cermak, Ed. (Erlbaum, Hillsdate, NJ, 1982), pp. 167-193.
4. E. Tulving, in The Cognitive Neurosciences, M. S. Gazzaniga, Ed. (MIT Press, Cambridge, MA, 1995), pp. 839-847.
5. The youngest ages of amnesia-inducing injuries reported previously were 9 [F. B. Wood, I. S. Brown, R. H. Felton, Brain Cognit. 10, 76 (1989)] and 10 [A. L. Ostergaard, Neuropsychologia 25, 341 (1987)].
6. The youngest ages of amnesia-inducing injuries reported previously were 9 [F. B. Wood, I. S. Brown, R. H. Felton, Brain Cognit. 10, 76 (1989)] and 10 [A. L. Ostergaard, Neuropsychologia 25, 341 (1987)].
7. An alternative possibility had been that very early medial temporal damage might fail to produce amnesia because of the reorganizational capacity of the immature brain. Indeed, unimpaired memory has been reported in cases of bilateral dysgenesis of the temporal lobes [P. W. Nathan and M. C. Smith, J. Neurol. Neurosurg. Psychiatry 13, 191 (1950); C. Lang, S. Lehrl, W. Huk, ibid. 44, 626 (1981)]. The present findings, which indicate that amnesia results from bilateral hippocampal pathology, however early such damage occurs (from at least birth onward), rule out this possibility as well.
8. An alternative possibility had been that very early medial temporal damage might fail to produce amnesia because of the reorganizational capacity of the immature brain. Indeed, unimpaired memory has been reported in cases of bilateral dysgenesis of the temporal lobes [P. W. Nathan and M. C. Smith, J. Neurol. Neurosurg. Psychiatry 13, 191 (1950); C. Lang, S. Lehrl, W. Huk, ibid. 44, 626 (1981)]. The present findings, which indicate that amnesia results from bilateral hippocampal pathology, however early such damage occurs (from at least birth onward), rule out this possibility as well.
9. Informed consent was obtained from all three patients after they had been given a description of the studies to be carried out.
10. D. Wechsler, J. Psychol. 19, 87 (1945).
11. J. L. Talley, Children's Auditory Verbal Learning Test (Psychological Assessment Research, Odessa, FL, 1993).
12. P. A. Osterreith, Arch. Psychol. 30, 306 (1944).
13. E. Isaacs and F. Vargha-Khadem, Br. J. Dev. Psychol. 7, 377 (1989).
14. A. Sunderland, J. E. Harris, A. D. Baddeley, J. Verb. Learn. Verb. Behav. 22, 341 (1983). The questionnaire asked the parents to rate how often their child forgets each of 29 prototypical daily events on a scale ranging from A (has not occurred in the past 3 months) to I (occurs more than once a day). The vast majority of the ratings by all three sets of parents were either G (more than once a week), H (about once a day), or I (more than once a day).
15. B. Wilson, J. Cockbum, A. Baddeley, Rivermead Behavioural Memory Test (Thames Valley Test Co., Reading, UK, 1985).
16. 1H MRS data from 2 by 2 by 2 cm cubes centered on the medial portions of the right and left temporal lobes, using a 90°-180°-180° spin-echo technique (TR, 1600 ms; TE, 135 ms) as described previously (16). Signal intensities at 2.0 parts per million (ppm) [primarily N-acetylaspartate (NAA)], 3.0 ppm [creatine + phosphocreatine (Cr)], and 3.2 ppm [choline-containing compounds (Cho)] were measured from the peak areas by integration and corrected for differences in radio-frequency coil loading among the different individuals. The data are presented in the form of the signal-intensity ratio NAA/ (Cho + Cr), which provides a simple index of spectral abnormality, with values below 0.72 (the lower limit of the 95% reference range) being indicative of neuronal loss or damage or astrocytosis, or both (76); J. H. Cross et al., Ann. Neurol. 39, 107 (1996); D. G. Gadian et al., Neurology 46, 974 (1996)]. The positioning and volume of the selected regions are such that the hippocampi in the amnesic subjects make only a minor contribution to the observed signal intensities.
17. 1H MRS data from 2 by 2 by 2 cm cubes centered on the medial portions of the right and left temporal lobes, using a 90°-180°-180° spin-echo technique (TR, 1600 ms; TE, 135 ms) as described previously (16). Signal intensities at 2.0 parts per million (ppm) [primarily N-acetylaspartate (NAA)], 3.0 ppm [creatine + phosphocreatine (Cr)], and 3.2 ppm [choline-containing compounds (Cho)] were measured from the peak areas by integration and corrected for differences in radio-frequency coil loading among the different individuals. The data are presented in the form of the signal-intensity ratio NAA/ (Cho + Cr), which provides a simple index of spectral abnormality, with values below 0.72 (the lower limit of the 95% reference range) being indicative of neuronal loss or damage or astrocytosis, or both (76); J. H. Cross et al., Ann. Neurol. 39, 107 (1996); D. G. Gadian et al., Neurology 46, 974 (1996)]. The positioning and volume of the selected regions are such that the hippocampi in the amnesic subjects make only a minor contribution to the observed signal intensities.
18. 1H MRS data from 2 by 2 by 2 cm cubes centered on the medial portions of the right and left temporal lobes, using a 90°-180°-180° spin-echo technique (TR, 1600 ms; TE, 135 ms) as described previously (16). Signal intensities at 2.0 parts per million (ppm) [primarily N-acetylaspartate (NAA)], 3.0 ppm [creatine + phosphocreatine (Cr)], and 3.2 ppm [choline-containing compounds (Cho)] were measured from the peak areas by integration and corrected for differences in radio-frequency coil loading among the different individuals. The data are presented in the form of the signal-intensity ratio NAA/ (Cho + Cr), which provides a simple index of spectral abnormality, with values below 0.72 (the lower limit of the 95% reference range) being indicative of neuronal loss or damage or astrocytosis, or both (76); J. H. Cross et al., Ann. Neurol. 39, 107 (1996); D. G. Gadian et al., Neurology 46, 974 (1996)]. The positioning and volume of the selected regions are such that the hippocampi in the amnesic subjects make only a minor contribution to the observed signal intensities.
19. W. Van Paesschen, A. Connelly, M. D. King, G. D. Jackson, J. S. Duncan, Ann. Neurol. 41, 41 (1997).
20. G. D. Jackson, A. Connelly, J. S. Duncan, R. A. Grunewald, D. G. Gadian, Neurology 43, 1793 (1993).
21. A. Connelly, G. D. Jackson, J. S. Duncan, M. D. King, D. G. Gadian, ibid. 44, 1411 (1994).
22. G. A. Press, D. G. Amaral, L. R. Squire, Nature 341, 54 (1989); C. R. Jack et al., Radiology 175, 423 (1990).
23. G. A. Press, D. G. Amaral, L. R. Squire, Nature 341, 54 (1989); C. R. Jack et al., Radiology 175, 423 (1990).
24. Although our MRS findings suggest that the bilateral temporal-lobe pathology is focal rather than widespread, the low spatial resolution of the technique does not preclude the possibility that the pathology extended into the underlying perirhinal and entorhinal cortices, an area known to be critical for many types of memory formation in monkeys [M. Meunier, J. Bachevalier, M. Mishkin, E. A. Murray, J. Neurosci. 13, 5418 (1993); B. J. Spiegler and M. Mishkin, Soc. Neurosci. Abstr. 5, 323 (1979); E. A. Murray and D. Gaffan, ibid. 19, 438 (1993); S. Goulet and E. A. Murray, ibid. 21, 1446 (1995)]. We therefore carefully examined the MRI 3D data sets for any evidence of structural abnormality in this specific region as well as in other regions throughout the brain. Although the possibility of undetected damage in the perirhinal and entorhinal regions or elsewhere in the three patients cannot be excluded, the only visible abnormalities noted outside the hippocampal formation were in Beth's scans, in which there was an increase in T2-weighted signal intensity in the periventricular and particularly peritrigonal white matter, accompanied by a marked loss of white matter bulk and thinning of the corpus callosum. These extrahippocampal abnormalities could well be the cause of Beth's impaired intellectual development, described earlier. No extrahippocampal abnormalities were detectable in the other two patients. Thus, as far as we could determine from the magnetic resonance findings, the only abnormality common to all three patients was bilateral pathology of the hippocampal formation.
25. Although our MRS findings suggest that the bilateral temporal-lobe pathology is focal rather than widespread, the low spatial resolution of the technique does not preclude the possibility that the pathology extended into the underlying perirhinal and entorhinal cortices, an area known to be critical for many types of memory formation in monkeys [M. Meunier, J. Bachevalier, M. Mishkin, E. A. Murray, J. Neurosci. 13, 5418 (1993); B. J. Spiegler and M. Mishkin, Soc. Neurosci. Abstr. 5, 323 (1979); E. A. Murray and D. Gaffan, ibid. 19, 438 (1993); S. Goulet and E. A. Murray, ibid. 21, 1446 (1995)]. We therefore carefully examined the MRI 3D data sets for any evidence of structural abnormality in this specific region as well as in other regions throughout the brain. Although the possibility of undetected damage in the perirhinal and entorhinal regions or elsewhere in the three patients cannot be excluded, the only visible abnormalities noted outside the hippocampal formation were in Beth's scans, in which there was an increase in T2-weighted signal intensity in the periventricular and particularly peritrigonal white matter, accompanied by a marked loss of white matter bulk and thinning of the corpus callosum. These extrahippocampal abnormalities could well be the cause of Beth's impaired intellectual development, described earlier. No extrahippocampal abnormalities were detectable in the other two patients. Thus, as far as we could determine from the magnetic resonance findings, the only abnormality common to all three patients was bilateral pathology of the hippocampal formation.
26. Although our MRS findings suggest that the bilateral temporal-lobe pathology is focal rather than widespread, the low spatial resolution of the technique does not preclude the possibility that the pathology extended into the underlying perirhinal and entorhinal cortices, an area known to be critical for many types of memory formation in monkeys [M. Meunier, J. Bachevalier, M. Mishkin, E. A. Murray, J. Neurosci. 13, 5418 (1993); B. J. Spiegler and M. Mishkin, Soc. Neurosci. Abstr. 5, 323 (1979); E. A. Murray and D. Gaffan, ibid. 19, 438 (1993); S. Goulet and E. A. Murray, ibid. 21, 1446 (1995)]. We therefore carefully examined the MRI 3D data sets for any evidence of structural abnormality in this specific region as well as in other regions throughout the brain. Although the possibility of undetected damage in the perirhinal and entorhinal regions or elsewhere in the three patients cannot be excluded, the only visible abnormalities noted outside the hippocampal formation were in Beth's scans, in which there was an increase in T2-weighted signal intensity in the periventricular and particularly peritrigonal white matter, accompanied by a marked loss of white matter bulk and thinning of the corpus callosum. These extrahippocampal abnormalities could well be the cause of Beth's impaired intellectual development, described earlier. No extrahippocampal abnormalities were detectable in the other two patients. Thus, as far as we could determine from the magnetic resonance findings, the only abnormality common to all three patients was bilateral pathology of the hippocampal formation.
27. Although our MRS findings suggest that the bilateral temporal-lobe pathology is focal rather than widespread, the low spatial resolution of the technique does not preclude the possibility that the pathology extended into the underlying perirhinal and entorhinal cortices, an area known to be critical for many types of memory formation in monkeys [M. Meunier, J. Bachevalier, M. Mishkin, E. A. Murray, J. Neurosci. 13, 5418 (1993); B. J. Spiegler and M. Mishkin, Soc. Neurosci. Abstr. 5, 323 (1979); E. A. Murray and D. Gaffan, ibid. 19, 438 (1993); S. Goulet and E. A. Murray, ibid. 21, 1446 (1995)]. We therefore carefully examined the MRI 3D data sets for any evidence of structural abnormality in this specific region as well as in other regions throughout the brain. Although the possibility of undetected damage in the perirhinal and entorhinal regions or elsewhere in the three patients cannot be excluded, the only visible abnormalities noted outside the hippocampal formation were in Beth's scans, in which there was an increase in T2-weighted signal intensity in the periventricular and particularly peritrigonal white matter, accompanied by a marked loss of white matter bulk and thinning of the corpus callosum. These extrahippocampal abnormalities could well be the cause of Beth's impaired intellectual development, described earlier. No extrahippocampal abnormalities were detectable in the other two patients. Thus, as far as we could determine from the magnetic resonance findings, the only abnormality common to all three patients was bilateral pathology of the hippocampal formation.
28. J. Rust, S. Golombok, G. Trickey, Wechsler Objective Reading Dimensions Test (Psychological Corporation, Sidcup, UK, 1993).
29. Like the tests used in monkeys (18), the tests for humans were designed with memory loads (that is, number of items and length of delays) that exceeded online working memory capacity so as to tap memory storage ability. It should be emphasized that these tests are not considered to be selective measures of either episodic or semantic learning [see (28)].
30. Photos of 20 different female faces were presented in succession on the computer monitor, with each one paired with a recording of that person's voice uttering the phrase "I am sorry I am not able to meet you, but I hope you will remember me." After presentation of the 20 face-voice pairs, one of the recorded voices uttering the same phrase was repeated while two faces were presented on the monitor: the correct face for that particular voice, together with an incorrect one associated with a different voice. The individual responded by touching one of the faces, which produced an indication of "correct" or "incorrect" on the screen, depending on the choice. Such choice tests for all 20 face-voice pairs completed a trial, and the trials were repeated until the individual achieved at least 18 correct responses in one trial or had received a maximum of 10 trials. Each of the normal control individuals learned these associations rapidly (mean, 1.8 trials; range, 1 to 5 trials) and accumulated very few errors (mean, 5.8 errors; range, 1 to 23 errors). By contrast, among the amnesic patients, only Jon performed as well as the poorest control individual, requiring five trials to learn and making 20 errors; Kate learned the associations in six trials (31 errors), and Beth was unable to reach the criterion in the 10-trial limit, making 57 errors. The group difference was significant at P < 0.05 for both trials and errors.
31. Pictures of 20 different objects were presented in succession on the right side of the computer monitor, with each object paired with a different circle in an irregular array of 40 circles located in the center of the monitor; the paired circle was illuminated and the object then appeared within it. After presentation of the 20 object-place pairs, one of the previously lit circles was relit, while two objects were presented on the right: the correct object for that particular circle in the array, together with an incorrect one associated with a different circle. The individual responded by touching one of the objects, which produced a voice-recorded "correct" or "incorrect," depending on the choice. Such choice tests for all 20 object-place pairs completed a trial, and the trials were repeated until the individual achieved at least 18 correct responses in one trial or had received a maximum of 10 trials. Each of the normal individuals learned these object-place associations rapidly (mean, 1.7 trials; range, 1 to 5 trials) and accumulated very few errors (mean, 4.2 errors; range, 0 to 19 errors). By contrast, none of the amnesic patients learned within the 10-trial limit, each one barely exceeding chance performance 5 (Beth, 86 errors in 200 responses; Jon, 75 errors; Kate, 95 errors). The group differences in trials and errors were each significant at P < 0.01.
32. J. K. Parkinson, E. A. Murray, M. Mishkin, J. Neurosci. 8, 4159 (1988). How spatial memory and episodic memory are related to each other in humans and nonhuman primates remains a high-priority issue for research.
33. E. A. Murray and M. Mishkin, Soc. Neurosci. Abstr. 22, 281 (1996).
34. B. J. Spiegler and M. Mishkin, Behav. Brain Res. 3, 303 (1981).
35. E. A. Murray, D. Gaffan, M. Mishkin, J. Neurosci. 13, 4549 (1993).
36. E. A. Murray and M. Mishkin, Science 228, 604 (1985).
37. This proposal does not imply that the computerized recognition tasks are tests of semantic learning; however, neither are they tests of episodic learning. The distinction between the two forms of cognitive memory applies not to the moment that memories are first acquired but rather to their long-term content. At the moment that attended sensory information is encoded for initial storage, the information is necessarily in the form of an unfolding event or episode replete with context, including time, place, surroundings, and persons; how much of the information in that episode is retained in long-term memory is what determines whether the memory is properly classified as episodic or semantic.
38. W. S. Suzuki, Semin. Neurosci. 8, 3 (1996).
39. S. Zola-Morgan, L. R. Squire, D. G. Amaral, J. Neurosci. 6, 2950 (1986); N. L. Rempel-Clower, S. Zola-Morgan, L. R. Squire, Soc. Neurosci. Abstr. 20, 1075 (1994).
40. S. Zola-Morgan, L. R. Squire, D. G. Amaral, J. Neurosci. 6, 2950 (1986); N. L. Rempel-Clower, S. Zola-Morgan, L. R. Squire, Soc. Neurosci. Abstr. 20, 1075 (1994).
41. W. B. Scoville and B. Milner, J. Neurol. Neurosurg. Psychiatry 20, 11 (1957).
42. We thank A. Incisa della Rocchetta for setting up the computerized tests, A. Incisa della Rocchetta and M.-C. Jones for testing some of the control individuals, and K. Pettigrew for statistical advice. Supported by the Wellcome Trust and Action Research.