The Neural Substrates
of Religious Experience
Jeffrey L. Saver, M.D.
John Rabin, M.D.
Religious experience is brain-based, like all human R eligious experience is brain-based. This should be
taken as an unexceptional claim. All human experi-
experience. Clues to the neural substrates of relig-
ence is brain-based, including scientific reasoning,
ious-numinous experience may be gleaned from
mathematical deduction, moral judgment, and artistic
tern porolimbic epilepsy, near-death experiences, creation, as well as religious states of mind. Determining
and hallucinogen ingestion. These brain disorders the neural substrates of any of these states does not
and conditions may produce depersonalization, automatically lessen or demean their spiritual signifi-
derealization, ecstasy, a sense of timelessness and cance.1 The external reality of religious percepts is nei-
ther confirmed nor disconfirmed by establishing brain
spaceless ness, and other experiences that foster
correlates of religious experience. Indeed, it has been
religious-numinous interpretation. Religious
argued that demonstrating the existence of a neural
delusions are an important subtype of delusional apparatus sustaining religious experience can reinforce
experience in schizophrenia, and mood-congruent belief because it provides evidence that a higher power
religious delusions are a feature of mania and has so constructed humans as to possess the capacity to
depression. The authors suggest a limbic marker experience the divine.2 For the behavioral neurologist
hypothesis for religious-mystical experience. The and neuropsychiatrist, the challenge is to delineate the
distinctive neural substrates of religious experience and
tern porolimbic system tags certain encounters with
their alteration in brain disorders.
external or internal stimuli as depersonalized, de- Investigation of the neural ground of religious expe-
realized, crucially important, harmonious, and/or rience is hampered by the absence of a widely accepted
joyous, prompting comprehension of these experi- animal correlate that would allow laboratory experi-
ences within a religious framework. mentation. Evidence with neuroanatomic import is
(The Journal of Neuropsychiatry and Clinical largely derived from clinical observations in patients
Neurosciences 1997; 9:498-510) with focal brain lesions, especially epileptic disorders,
and much of this is anecdotal. We will review data that
have been collected on religious experience in normal
individuals and in different neurologic and neuropsy-
chiatric syndromes. From this scattered literature, a pre-
From UCLA-Reed Neurologic Research Center, 710 Westwood Plaza,
Los Angeles, CA 90095. Send correspondence to Dr. Saver at the above
address.
Copyright © 1997 American Psychiatric Press, Inc.
498 VOLUME 9 #{149}
NUMBER 3 #{149}
SUMMER 1997

SAVER AND RABIN
liminary unifying model of the brain basis of religious that perception of the divine occurs not through the
experience may be constructed. operation of a distinctive sensory faculty, but through
the superimposition upon ordinary sensations of a
unique numinous-mystical feeling, a feeling of direct
CIRCUMSCRIPTION OF THE TOPIC awareness of a sacred or divine presence.8’9 Studies of
healthy individuals and neuropsychiatric populations
Most religious experience parallels ordinary experience. support this hypothesis and suggest a distinctive
The religious sentiments include religious joy, religious neurolimbic substrate.
love, religious fear, and religious awe. These religious
emotions are analogues of ordinary emotions of joy,
love, fear, and awe, differing not in their emotional RELIGIOUS EXPERIENCE IN
tone, but only in being directed to a religious object. NORMAL INDIVIDUALS
Their neural substrate is likely to contain nothing of a
specifically religious nature, but instead to rest upon the Surveys suggest that religious-numinous experiences
same limbic and subcortical networks that support non- are common in both children and adults, across differ-
religious joy, love, fear, and awe, directed by dorsolat- ent historical eras, and across all cultures. #{176}‘ In national
eral and orbital frontal cortices to religious rather than surveys in the United States, Britain, and Australia, 20%
nonreligious targets. The neural substrates of human to 49% of individuals report having personally had
emotionality have been extensively delineated.34 numinous experiences, and this figure rises to more
Similarly, religious language depends upon the cus- than 60% when in-depth interviews of randomly se-
tomary dominant-hemisphere perisylvian language lected individuals are conducted.8 Hardy and his col-
cortices for its production, differing only in taking sa- leagues8’6 identified eight major types of numinous
cred rather than nonsacred topics as linguistic themes. experience in British individuals, encompassing, in de-
Focal left-hemisphere lesions produce aphasia for relig- scending order of frequency, 1) a patterning of events
ious discourse that parallels aphasia for nonreligious in a person’s life that convinces him or her that in some
discourse, as common clinical experience attests. Pros- strange way they were meant to happen, 2) an aware-
ody and other emotional contributions to discourse of ness of the presence of God, 3) an awareness of receiving
the right hemisphere apply to both religious and non- help in answer to a prayer, 4) an awareness of being
religious themes, and nondominant-hemisphere looked after or guided by a presence not called God, 5)
lesions have been reported to produce parallel impair- an awareness of being in the presence of someone who
ments of religious and nonreligious emotional process- has died, 6) an awareness of a sacred presence in nature,
ing.5 Similarly, scholastic/talmudic reasoning is 7) an awareness of an evil presence, and 8) experiencing
ordinary reasoning applied to religious problems and is in an extraordinary way that all things are “One.” Stud-
undoubtedly mediated by the same neural networks in ies in identical and fraternal twin pairs raised apart
frontal and parietal multimodal association areas. suggest that genetic factors account for 50% of interim-
A first general observation, then, is that the neural dividual variance in religious interests and attitudes.’7
substrate for the preponderance of religious affect and The cross-cultural ubiquity of numinous experiences
cognition is the whole human brain, employing proc- and the heritability of religious dispositions argue
essing that is parallel, distributed, affective, and sym- strongly for a biologic basis, but fail to indicate the
bolic, with contributions of large-scale neurocognitive specific neural mechanisms involved. Clues to neural
networks subspecialized for linguistic, prosodic, logical, substrate must be gleaned from the sites of brain disor-
and affective processing. ders that provoke qualitatively similar experiences.
What might be peculiarly distinctive to religious ex-
perience would appear, on first inspection, to reside not
in the domains of affect, language, or cognition, but in EPILEPSY AND RELIGIOUS EXPERIENCE
perception. It is the direct sensory awareness of God or
the divine that is a quintessential mark of specifically Humanity has long recognized a direct link between
religious experience. There is, however, no identifiable epilepsy and the divine. The early Greeks viewed epi-
separate organ of religious perception. Accordingly, lepsy as a “sacred disease,” a visitation from the gods,
sensory apprehension of the divine is likely to be medi- until the notion of divine genesis of seizures was dis-
ated, at least in part, by the neural systems for ordinary pelled by Hippocrates.18 In the Medieval and early Ren-
tactile, visual, auditory, and olfactory perception. Wil- aissance periods, wide currency was given the biblical
liam James and others have suggested, and we concur, view that epileptic seizures are manifestations of de-
JOURNAL OF NEUROPSYCHIATRY 499

NEURAL SUBSTRATES OF RELIGIOUS EXPERIENCE
monic possession (Mark 9:14-29).’ Esquirol2#{176} in 1838 arousal with set cognitive schemata. Epileptic auras
and Morel2’ in 1860 recognized a heightened “religios- themselves are the most convincing evidence of the
ity” of epileptics, which they attributed to disability, insufficiency of attribution theory, instead demonstrat-
social isolation, and greater need for religious consola- ing the existence of distinct physiologic neural sub-
tion. The first reported conversion experience directly strates for several specific emotional states. Several
related to a seizure was noted by Howden, whose “psychic” auras, including depersonalization, dereali-
patient experienced being transported to Heaven dur- zation, dreamy states, autoscopy, and ecstasy, are par-
ing a fit. Spratlin3 in 1904 reported a religious aura in ticularly likely to engender religious interpretation and
52 of 1,325 patients with epilepsy (4%). In the early 20th experience, and merit detailed review.
century, Turner24 and others suggested that epileptics Among individuals with partial complex seizures, the
develop a distinctive interictal character, among the frequency of auras ranges from 23% to 83%, and up to
features of which is religious fervor. A substantial num- one-quarter of auras are psychic in content.2729 The most
ber of founders of major religions, prophets, and lead- common psychic or experiential ictal manifestations of
ing religious figures have been documented as having temporal lobe epilepsy are fear, d#{233}j#{224}
vu,jamais vu, mem-
or suggested to have epilepsy (Table 1). ory recall, and visual and auditory hallucinations.
Modern investigations of epilepsy-related religious Hughlings Jackson was among the first to identify
experience have been marked by contradiction and con- and characterize less common ictal “intellectual auras”
flict. This confusion in part reflects failure to distinguish (“dreamy states,” “cognitive auras”) in which the expe-
among epilepsy-related religious experiences of ictal, rience of the immediacy and liveliness of one’s own or
subacute postictal, and chronic interictal occurrence. external reality is altered.31 The intellectual auras in-
The phenomenology and neurobiology of each differ in clude depersonalization, derealization, and double con-
important aspects. sciousness. Depersonalization auras produce an
alteration or loss of the sense of one’s own reality, often
Ictal Events accompanied by a sense of detachment from others and
Ictal events of any type may be the subject of religious the environment, and of acting like an automaton. Dere-
or cosmological explanation. Seizures are paroxysmal, alization auras generate an alteration or loss of the sense
riveting, and unexpected-sudden intrusions of unan- of the reality of the external environment-for example,
ticipated and often extraordinary experience into the the feeling that the external surround is just a dream-
ordinary daily flow of consciousness. Patients who have and also are often associated with a sense of detach-
culturally acquired explanatory systems of a religious ment. Double consciousness (“mental diplopia”) auras
character naturally tend to interpret any ictal experi- create a simultaneous experience of persisting remnants
ence as possessing religious significance. Studies have of one’s normal consciousness and of a new quasi-
demonstrated that experiences that are personal, im- parasitical consciousness with a different perception of
portant, negative, and medical, like most seizures, are reality. Auras of depersonalization, derealization, and
particularly likely to be interpreted in a religious frame- dreamy state account for approximately one-quarter of
work.25 psychic auras.32
Some psychologists of religion have attempted to Well-documented localizations of spontaneous dis-
explain religious experience by employing classical at- charges or electrical stimulations producing intellectual
tribution theory. Experiments demonstrate that indi- auras are extremely rare. Available evidence, however,
viduals not only interpret, but also inwardly experience, suggests that mesolimbic structures, the hippocampus
the same physiologic stimuli in strikingly different ways and especially the amygdala, are likely the critical gen-
according to the cognitive expectations they carry. In erators of a feeling of unreality about the self or external
laboratory studies, identical, pharmacologically in- reality.33 The experience of unreality occurred in 9%
duced sympathetic arousals are variably labeled and are of patients with temporal lobe epilepsy in one series,
experienced across the spectrum of valence, from re- often accompanied by a sensation of fear.35 One patient
warding to distressful, according to the individual’s in Penfield’s series had a sense of “not being in this
cognitive set at onset.26 Attribution theory predicts that world” each time the first temporal gyrus was stimu-
religious individuals with epilepsy will often experience lated, reflecting either direct lateral temporal cortical
intrinsically neutral ictal physiologic events as having a excitation or rapid spread of afterdischarge to mesolim-
religious-numinous character. This mechanism may in- bic structures. One patient in Gloor and colleagues’
deed mediate some ictal religious experiences. more recent series37 repeatedly experienced a faraway
Human experience, however, is clearly not simply the feeling during stimulation of temporal mesolimbic
product of an interaction of nonspecific physiologic structures.
500 VOLUME 9 #{149}
NUMBER 3 #{149}
SUMMER 1997

SAVER AND RABIN
E
‘ -‘
0
a .1
I)
..
a
.0._t >. 0
.-0-
E e
0;
. asa
U’’0 aO. j.0’
U
a Oa a5iI.. za
Li
81
u .
u’ uEa .
0.
a -
a - .
0. ‘B B B
81 ‘
,j a..abc a 2 #{149}0a0 ‘‘ a
0a a . . U
a :0 a a
.0. #{149}E U
0.0 0
.0as
a0.0
aO .0 a a u
0 0
LI.0. , LLI LI LI U LI. 1.1,
,.,
0a
a a“ a a s
a .
0_ . . . .
a a a a a
so a
i;’
U
a . U.0
a a
#{149} 8
#{149} 8
0 a
.0
I.’
0.
-2
0 a.’ i.. i.
a a a a a a a
n. 2 0. 0. 0. 0.
C
-.
x
a . . U’8
0. 0 0 a.1 a a. a a<
> a n.
.0 ,
. .: 0.
2 a E 2
0 - B.
0 0. 0 0’
U W0. Li .2 u u u u LU
a 81
0. + + +
01 + + + + + + +
+
0.
01
#{149}0 c’.
a a- a’
4
.. a .-‘
01 a
I- :‘
a a
0’ a
.C
V 0 a aao o a a
0
C
a 0 020 C
01 0
=-.0 F-0
a #{149}-a a
C.
0 - a
a
a
5
a
a- II.
a
U
a) Ga
01 IL, .< LU U.
a
V
.
81
E
..c,,,.a
a
.0 0
Oa
a
a
a
.
.
081
a
.0a8 a
.‘.‘
0
0
-
.0
a
U
a
tJ- 2a0 :0
01 aa :0 a 8
500 aO E0 a 0
a
>s :00
C E.5 X 0 a
0 b00U= a
01 0
0 Oa:0aa.. .c
0
a a aU
01 0
‘ a
0
81 abO0 aa . a.0
0. U .0 ..03 0a .0
a
a
2_0m .0 .
‘a. .
a aaa 0 a00a
0 .0 o.
C 0.Q. 22 500 U aEij 2 2
.5a0 a ..‘,8J .0
0 0
:o0 U
a a0CU a
‘C i0> ,j
0. > a 0.tj
a
a Li
a
-
,-a U o Un - c8L)0
a
C
so
01
a
U00. .Oa:0,...E :- c a
U LU 0. < , > - < >
=
a -. r’)
8 a
Li
j ‘C
a’ 0 U 0
v-’ a .0
U . ._ 0
a .. a u a ‘
n.m ‘
‘? E . aS ‘< “ a. ‘
0
a . 0 c .
. 5 <
a 2
a 0. a
0 a . 0 a a 0
.E U- F-’-.’
.0 “U C U-’ a
a.- a .- .- .-
(I) #{163} c u <
JOURNAL OF NEUROPSYCHIATRY 501

NEURAL SUBSTRATES OF RELIGIOUS EXPERIENCE
__7 .
C
-.n
0 ,0 a-
U 0)
V
U aCa C
01 . .#{176}#{176}.‘ ‘ .. -
‘ nUa)a-0 a n.o C
E aa a
o a -
C 0gb .00
01 a
0 a C
.2
a
U
a 0 a 0
a.- N 0 . #{176}
a-
a-
a a a a a a- a a
0 . . 0 . . .& . >.‘
0) a,.. a a 0J aa... a
a N a a.0 .
a a- .E .9
a-a aa0
a- a
a a aa0 a
0.
na.
0. 0
a.0.’.’
0. ,..
a
0.
x xa xa x
#{149} o a5n. a .
n.ao) n.
S 5#{176}Ea 5 E.’-.
0
U U 0.X U U.. U
- .5. + + +
: + + + + + +
ga . + +
a
a a m
.0 0) 0
a a
01 > 0
0 a a 81aa a
a . N = ‘
om a81 .2
.0 C a 0.0’- 0.50
QJ ,-,a0 an
01 01 >0 0 .‘..‘=a
a- 0 U001 01a)
C. 010> 50 a -
o P.00 2a a a
U
1810 LU
a
U C
01 ‘-2. 0 ‘
E - oE” a
01 OE.0 ‘E C
GaO
:0a50 a-
a U
‘a-.0 o
‘ . aU 0Jo a
a -a
. ‘aaa
a- 01 0. > a Ca. a- Ca .-.
a IC a)
aai
V.0 a01 .0.’.’ ba 0.0
C.a0G o.
a _ ‘ 00 Ca-a O5 .5> -
a C a)0
#{149}.0 C500.
._>.-C a .0”c n
C a>5a
0 a-
. 5050a0aCa aa
0 Ga- Ca a’-
a-C
aa
‘C 0. .0Ca 0UaC
a
0 i2 a-aa a a>
a)0
0
V
a.0__
0.
o25a.
01 0
U 0 an
Ox
U
cia 0. 0. cia
0.
a
.0 0 a.-., X 0.01
a
C a .C-’ 1181
a
a....,
.0 ‘B
0 .C oo U
oars C
0
a ,- -.‘
F- 0. LI.
502 VOLUME 9 #{149}
NUMBER 3 #{149}
SUMMER 1997